PREDICTORS OF PROBLEM BEHAVIOURS
INTHE
STUDENT POPULATION SERVED
BY THE SASKATOON TRIBAL COUNCIL
A Thesis Submitted to the College of
Graduate Studies and Research
in Partial Fulfilment of the Requirements
for the Degree of Master of Education
in the Department for the Education of
Exceptional Children
University of Saskatchewan
Saskatoon
by
David Brian Mykota
March, 1996
© Copyright David B. Mykota, 1996. All rights reserved .
. �,__---
-.-� '--"��-"------The authour has agreed that the Library, University of Saskatchewan, may make this thesis freely available for inspection. Moreover, the authour has agreed that permission may be granted by the professor or professors who supervised the thesis work recorded herein or, in their absence, by the Head of the Department or the Dean of the College in which the thesis work was done. It is understood that due recognition will be given to the authour and the
University of Saskatchewan in any use of the material in this thesis. Copying or publication or any use of the thesis for financial gain without approval by the
University of Saskatchewan and the authour's written permission is prohibited.
Requests for permission to copy or to make other use of material in this thesis in whole or in part should be addressed to:
Head of the Department
for the Education of Exceptional Children
University of Saskatchewan
Saskatoon, Saskatchewan
Canada. S7N OWO
.. -�
..... i.II ABSTRACT
Children and adolescents at risk for school failure have become the focus of increased attention because the severity of problems facing them has been intensifying. Teachers report that at-risk children and adolescents in the classroom are not only effecting their own educational opportunities but also those of their peers. Furthermore, in the area of education services for First
Nation communities, biases in the identification of at-risk students has been suggested to exist. In order for effective remediation to occur, it is necessary to examine those processes which place children and adolescents at-risk. As problem behaviours have been identified as placing children and adolescents at risk for failure in school or life, those variables which influence the development of problem behaviours need be understood.
Thus the purpose of the present study is to investigate the predictors of problem behaviours which place the student population served by the
Saskatoon Tribal Council at risk for failure in school. Specifically, the study examines the predictors of problem behaviours as measured by the Teacher
Rating Scale of the Behaviour Rating Profile-2nd Edition with the variables sex, age group, school, nonverbal reasoning ability as measured by the Matrix
Analogies Test-Short Form, and academic performance as measured by the teacher completed Academic Performance Rating Scale.
In determining the strength of the relationship between the variables problem behaviours, academic performance, non-verbal reasoning ability, age group, sex, and school correlates are presented. Further analysis of the data
ii
---- using multiple linear regression, with problem behaviours as the criterion variable and non-verbal reasoning ability, academic performance, school, sex,
and age group as the predictor variables are reported. The present study also
determined if there is a significant difference in problem behaviours for the
student population served by the Saskatoon Tribal Council as based on age group (6 to 9 years of age, 10 to 13 years of age, and 14 to 18 years of age), sex .
(male and female), and school (Band or provincial).
The results of the study indicate that because academic performance continues to be such a strong predictor of problem behaviours, underscores the importance of the school as an ameliorative factor in the prevention and treatment of behavioural problems. As well, because females 10 to 13 years of age attending provincial schools are rated with fewer problem behaviours than their counterparts attending Band schools, or those in other age groups, is
indicative of the need for the appropriate allocation of resources for children and youth at-risk. Given these findings, the present study recognizes the significance of intervention as a preventive measure for all age groups and the responsibility of federal and provincial funding agencies to continue to make available to First
Nation communities resources appropriate for the development of educational
services for students at-risk.
iii r .,
ACKNOWLEDGEMENTS
I would like to acknowledge the permisslon received by the Saskatoon
Tribal Council and Gordon Lobe to conduct the present study. As well, Lorna
Robert's diligent gathering of the data and her own thesis research pertaining to
the Saskatoon Tribal Council study on exceptionalities is duely acknowledged.
I would like to thank my external examiner Dr. Fred Reekie for the insight
provided and my committee members Dr. Barbara Bloom and Dr. Don
Saklofske for their guidance and support. In conclusion, I would like to
acknowledge my advisor Dr. Vicki Schwean for offering me the opportunity to
undertake the present study and for her continued commitment to research in
exceptionalities.
iv DEDICATION
This thesis is dedicated to the children and youth who participated in the
- Saskatoon Tribal Council study on exceptionalities.
v TABLE OF CONTENTS
PERMISSION TO USE . ABSTRACT...... ii
ACKNOWLEDGEMENTS...... iv
DEDICATION...... v
TABLE OF CONTENTS.... vi
LIST OF FIGURES...... xi
LIST OF TABLES... xii
1 INTRODUCTION...... 1 1 .1 Purpose...... 3 2 LITERATURE REVIEW...... 7 2.1 Demographics of Aboriginal People in Canada...... 9
2. 1 � 1 Population Distribution...... 9 2.1.2 Birth Rates...... 1 2 2.1.3 Dependency Ratio... 15 2.1.4 Living Conditions...... 16 2.1.5 Education...... 1 8
2.1.6 Economic Status...... 20 2.1.7 Unemployment and Labour Force Participation Rates...... 21 2.1.8 Language...... 24 2.1.9 Summary... 24 2.2 Mental Health in Aboriginal Communities...... 25 2.2.1 Definitional Criteria...... 26
2.2.2 Indicators of Mental Health. 27
' 2.2.2.1 Mortality 28 2.2.2.2 Alcohol and Substance Abuse...... 30 2.2.2.3 Family Violence and Child
Sexual Abuse . 31
2.2.2.4 Summary . 31 2.2.3 Epidemiological Studies...... 33
2.2.3.1 Ontario Child Health Study . 34 2.2.3.2 Ontario Child Health Study: Findings and Limitations...... 35 vi
" " _�;.c·: � :�,.-':"(,. �.¥• .":": �Ir- -''I' -",," 2.2.4 Native American Child and Adolescent Mental
Health...... 40 2.2.4.1 Statistical Overview...... 41 2.2.4.2 American Indian Health Survey...... 42 2.3 Variables Affecting Behaviour in Children and Adolescents...... 44 2.3.1 Special Education Students and Delinquency...... 45 2.3.2 Protective Effects of Intelligence on Delinquency...... 46 2.3.3 Familial Processes. 48 2.3.4 Familial Factors, School Behaviour, and Learning Disabled Students...... 51 2.3.5 Socio-economic Status, Behaviour, and 52 Intelligence :...... 2.3.6 Summary...... 54 2.4 Psychopathology Across Cultures...... 55
2.4. 1 Culture Bound Syndromes...... 56 2.4.2 Psychological Change and Acculturation...... 57 2.4.3 Acculturative Stress...... 58 2.4.4 Somatization and Social Response to
Distress...... 60
2.4.5 Deviance...... 61 2.4.5.1 Culturally Labelled Deviance...... 61 2.4.5.2 Societal Integration and Deviance...... 62 2.4.5.3 Stigmitization...... 63 2.4.6 Schizophrenia...... 64 2.4.7 Depression...... 65 2.5 Cross Cultural Assessment...... 66
2.5.1 Formal Assessment...... 67
2.5.2 Emics and Etics...... 68 2.5.2.1 The Combined Emic-Etic Approach...... 70 2.5.3 Psychometric Equivalence and Cross-Cultural Measurement Strategies...... 71
vii
'"- ..
. <._. ------"._ ••_ i,:�.,,,-) �_____ ,......
2.5.3.1 Direct Comparison and Crude Translation...... 71 2.5.3.2 Regression Methods...... 72 2.5.3.3 Co-scoring Methods...... 72 2.5.3.4 Response Pattern Method..... 72 2.5.3.5 Translation Techniques...... 73 2.5.3.6 Internal Structure Congruence...... 73 2.5.3.7 Nomological Network and Validation..... 75 2.5.3.8 Multi Strategy Approach...... 75 2.5.4 Assessment and Aboriginal North Americans..... 76 2.5.4.1 Developmental Change and Assessment...... 76
2.5.4.2 Deficit and Assessment..... 77 2.5.5 Testing and Aboriginal Students...... 78 2.5.5.1 Ability and Achievement Measurement . in Aboriginal Populations...... 79 2.5.5.2 Behaviour and Personality Measurement in Aboriginal Populations...... 82 2.5.6 Culturally Appropriate Assessment Practices..... 83 2.5.7 Methodological Concerns. 88 2.5.8 Summary...... 89 2.6 The Present Study...... 90 2.6.1 Research Questions...... 92
2.6.1.1 Research Questions # 1 and # 2... 92
2.6.1.1 .1 Rationale...... 92
2.6.1.1.2 Question#1...... 92
2.6.1.2.3 Question#2...... 92
2.6.1.2 ResearchQuestion#3...... 93 2.6.1.2.1 Rationale...... 93 2.6.1.2.2 Question#3...... 93 2.6.2 Assumptions...... 94 2.6.3 Umitations...... 94
2.6.4 Definitions...... 94
3 METHODS AND PROCEDURES...... 97 3.0 SubjectSample...... 97 viii 3.1 Preliminary Protocol...... 98 3.2 Data Collection ·...... 99 3.3 Instrumentation...... 100 3.3.1 Teacher Ratings...... 101 3.3.2 Behaviour Rating Profile...... 102 3.3.3 Academic Performance Rating Scale...... 1 05 3.3.4 Matrix Analogies Test-Short Form.. 107 3.4 Demographics...... 109 3.5 DataAnalysis 109 4 RESULTS...... 112
4.1 Overview...... 11 2 4.2 Descriptive Statistics...... 114 4.3 Evaluating Assumptions of Normality...... 115 4.4 Research Question Results...... 117 4.4.1 ResearchQuestion#1...... 117
4.4.2 Results and'Discussion...... 11 8
4.4.3 ResearchQuestion#2...... 119
4.4.4 Results and Discussion...... 119
4.4.5 Research Question#3...... 124
4.4.6 Results and Discussion...... 125 4.5 Summary of Results...... 130 5 DISCUSSION...... 132 5.1 Overview of Findings...... 132 5.2 Findings...... 134 5.3 Educationallmplications...... 140 5.4 Research Implications...... 143 5.4. 1 Culturally Sensitive Research...... 143 5.4.2 Risk Factors and Resilience...... 144 5.4.2.1 Longitudinal Studies...... 1 44 5.4.2.2 Developmental Perspectives...... 145 5.4.2.3 Situation Specificity...... 146 5.4.2.4 Multiple Regression and Replication...... 146 5.4.2.5 Gender Bias 147 5.5 Conclusion 148
REFERENCES...... 1 50
ix APPENDICES...... 164
Appendix 1
Saskatoon Tribal Council Letter of Permission...... 165
Appendix 2
Academic Performance Rating Scale...... 167 Appendix 3
Stepwise Multiple Regression for Non transformed Data...... 170
x LIST OF FIGURES
Figure 2.1 Percentage Growth Rate of Status Indian, Inuit, and Total Canadian Populations...... 1 0
Figure 2.2 Proportion of Person's of Native Origin and Total Canadian Population by Region...... 11
Figure 2.3 On-Reserve, Off-Reserve, and Total Canadian Population by Region...... 12
Figure 2.4 Age Structure of Status Indian and Total Canadian Populations...... 13
Figure 2.5 Crowded Dwellings: On-Reserve, Off-Reserve, Inuit, and Total Canadian Populations...... 16
Figure 2.6 Single Parent Families: On-Reserve, Off-Reserve, Inuit, and Total Canadian Populations...... 17
Figure 2.7 Children in Care for On-Reserve and Total Canadian Population, 1981-1987...... 18
Figure 2.8 On-Reserve, Off-Reserve, and Total Canadian Populations with less than Grade 9...... 19
Figure 2.9 On-Reserve, Off-Reserve, and Total Canadian Population with High School Diploma...... 20
Figure 2.10 Income Level of Families On-Reserve, Off-Reserve, and Total Canadian Population...... 21
Figure 2.11 Unemployment Rate of On-Reserve, Off-Reserve, Inuit, and Total Canadian Population...... 22
Figure 2.12 Labour Force Activity for On-Reserve, Off-Reserve, Inuit, and Canadian Populations...... 23
Figure 2.13 Schematic Operationalization of Emics and Etics...... 71
Figure 4.1 Interaction Effect (Males) for Age Group and School. 128
Figure 4.2 Interaction Effect (Females) for Age Group and School...... 1 28
Figure 5.1 Correlation Model for Predictors of Problem
8ehaviours , 134
xi LIST OF TABLES
Table 2.1 A Cross Cultural Assessment Model...... 87
Table 3.1 Sample byAge...... 99
Table 3.2 Sample bylnstrument...... 101
Table 4.1 Sample by Age and by Cognitive, Academic, and Behavioural Measure...... 114
Table 4.2 Age Group by School and Sex , 115 Table 4.3 Distribution Statistics by Instrument for Total Sample...... 116
Table 4.4 Results of Data Transformation.. 117
Table 4.5 Correlations by Variable for Total Sample...... 118
Table 4.6 Stepwise Multiple Regression Predictions of Problem Behaviours (TRS of the BRP�2)...... 120 Table 4.7 Collinearity Diagnostics...... 122
Table 4.8 Outliers at 3 Standard Deviations...... 124
Table 4.9 Descriptive Statistics for the Teacher Rating Scale of the BRP�2...... 125
Table 4.10 Analysis of Variance Teacher Rating Scale of the'
BRP-2 by Age Group, Sex, and School...... 126
Table 4.11 Analysis of Variance Teacher Rating Scale of the
BRP-2 (Males) by Age Group and SchooL...... 127
Table 4.12 Analysis of Variance Teacher Rating Scale of the
BRP-2 (Females) by Age Group and SchooL...... 127
Table 4.13 Analysis of Variance Teacher Rating Scale of the BRP-2 (Provincial Schools) by Age Group and Sex 130
Table 4.14 Analysis of Variance Teacher Rating Scale of the BRP-2 (Band Schools) by Age Group and Sex...... 130
xii
.)M "'�'II
..•It ...... ,... �------..----..------CHAPTER ONE
INTRODUCTION
The study of the relationship between psychiatric disorders and the demographic, cognitive, and academic correlates among children and youth was pioneered in the Isle of Wright surveys conducted by Rutter during the mid
1960's. Rutter's work on the Isle of Wright is of significance because it was the first large scale epidemiological study of childhood and adolescent educational, psychiatric, and physical disorders completed. Since the Isle of Wright surveys, a number of epidemiological studies have been undertaken to help ascertain what are the at-risk and protective factors in child development, future delinquency, and quality of adult adaptation to stressful life events. Most noteworthy is the Kauai longitudinal study (Werner & Smith, 1992) which continues to monitor the impact of biological and psychosocial risk factors, stressful life events, and protective factors on the development of a birth cohort into adulthood.
Within Canada, the Ontario Child Health Study (OCHS), conducted in 1983 in the Province of Ontario, is the most comprehensive study of child and adolescent epidemiology. The OCHS (Offord, Boyle, & Racine, 1989, 1991) sought to estimate the prevalence of psychiatric disorders within child and youth populations while at the same time exploring the correlates of the disorders.
Although the OCHS was extensive in its survey sample, three groups of children and youth between 6 to 16 years of age were excluded. One of these groups represented First Nation children living on their reserve. This occurred
because the researchers recognized the lack of instrument standardization as
1 problematic in an epidemiological study which included Aboriginal children
(Armstrong, 1993). As a result of this shortcoming, studies are being conducted
in some Aboriginal communities by the same researchers to rectify earlier
concerns (Kirmayer, 1994a). The need for these studies is made manifest when
it is realized that Native adolescents are at greater risk for the development of
emotional disorders (Beiser & Attneave, 1982; Health and Welfare Canada
Medical Services Branch, 1992; Kirmayer, 1994a, 1994b; Report of the Steering Committee Medical Services Branch, 1991). This is not surprising considering
the stressors placed on youth experiencing the effects of cultures in contact.
Even though there is epidemiological data available for Canadian child and
youth psychopathology, information pertaining to Aboriginal populations is
scant (Kirmayer, 1994a). This type of research is imperative if culturally
sensitive educational interventions are to be developed which take into account
the special needs and challenges facing Aboriginal children and adolescents
today.
Similarly, in the area of education services, when First Nations and the
federal government use the criteria for high cost special education services as
dictated by provincial formulas, biases in the identification of special needs First
Nation students have been suggested to exist. Specifically, concerns regarding
the instrumentation and measurement techniques advocated by the provincial
formula are circumspect when applied to First Nation populations (Schwean &
Saklofske, 1994). To rectify this shortfall in epidemiological studies of
exceptionality among First Nation children and to assist in the development of
school based treatments and interventions, the Saskatoon Tribal Council (STC)
initiated a research project under the auspices of University of Saskatchewan
faculty: Dr. V.L. Schwean, Department for the Education of Exceptional
Children; Dr. D.H. Saklofske, Department of Educational Psychology; and
2 ,-,,: I t conducted by Ms. L. Robert, a graduate student in the Department for the
Education of Exceptional Children.
1.1 Purpose
The purpose of the present study is to determine the predictors of problem
behaviours which would place the child and youth population of the Saskatoon
Tribal Council at-risk. The present study is descriptive in nature and will draw on the Saskatoon Tribal Council special education research undertaken by Dr. V.
L. Schwean, Department for the Education of Exceptional Children, Dr. D. H.
Saklofske, Department of Educational Psychology, and Ms L. Robert,
Department for the Education of Exceptional Children. Permission to conduct an
analysis of the screening instruments and data obtained from the Saskatoon
Tribal Council's special education study was received through Gordon Lobe,
Superintendent of Education for the Saskatoon Tribal Council (personal
communication, February 14, 1994), see Appendix 1.
Longitudinal and cross sectional research in child psychopathology has
investigated both at-risk and protective factors in children and adolescents who
are vulnerable to a disorder and susceptible to biological and psychosocial risk factors (Offord et aI., 1989, 1991; Rutter, 1989; Werner & Smith, 1992). Briefly
stated:
Resilience and protective factors are the positive counterparts to both vulnerability, which denotes an individual's susceptibility to a disorder, and risk factors, which are biological or psychosocial hazards that increase the likelihood of a negative developmental outcome in a group of people. (Werner & Smith, p. 3)
In the present study, at-risk will refer to those students who are likely to fail in
life or school due to a number of factors that include economic, social, and
physical variables which act as barriers to their normal development (Frymier &
Gansneder, 1989). The term at-risk is not intended to be pejorative, rather
3 children at risk have become the focus of increased attention because the severity of the problems facing at-risk children has been intensifying (Planning and Evaluation Branch, Saskatchewan Education, Training and Employment,
1994). Teachers in Saskatchewan report that the prevalence of at-risk children in the classrooms is increasing and that these children are not only effecting their own educational opportunities but also those of their peers (Planning and
Evaluation Branch, Saskatchewan Education, Training and Employment). In order for at-risk children to lead productive lives, it is necessary to examine the social, emotional, academic, and cognitive variables to determine what factors impede at-risk children's ability to succeed in school, lead productive lives, and contribute fully to society.
In the Kauai longitudinal study conducted by Werner and Smith (1992), which encompassed both quantitative and qualitative research methodologies, it was found that in determining predictors of delinquency and the likelihood of a negative developmental outcome in both the male and female child and adolescent populations, a low IQ (<90) coupled with behaviour and learning problems for girls and boys, respectively, were the most powerful predictors of future delinquency. Of those who went on to become persistent offenders in youth, a significant higher proportion were described by their classroom teachers as having uncontrollable emotional problems and being extremely irritable (Werner & Smith). Other risk factors which were predictors of a negative psychosocial developmental outcome were low socio-economic status, a low level of education for the mother, and a congenital defect or handicap in the child (Werner & Smith).
In a recent study of educational risk and resilience among African-American youth in Atlanta and New York, Connel, Spencer, and Aber (1994) found:
.,. that disaffected behavior in low-income African-American youth can lessen parental involvement, which in turn contributes to negative appraisals of self that exacerbate disaffected patterns of action [emotional 4
r I and behavioral engagement] and contribute to negative educational outcomes [low attendance, low test scores, low grades, suspension, and retention]. In order to avert these destructive transactional processes, adults, including but not restricted to parents, will have to reinterpret disaffected behavior on the part of youth and find ways to increase (not decrease) their involvement with these youth. (p. 504)
The implication for teachers who encounter the disaffected behaviour of youth which contributes to their negative educational outcomes is the need to facilitate . and maintain the educational involvement of these youth rather than allowing destructive processes to continue.
In terms of the Aboriginal child and youth population in Canada, similiar studies need to be undertaken if Aboriginal children are to succeed in the classroom and in life. Since Werner and Smith (1992) have validated classroom teachers as viable observers of problem behaviours which can be predictive of delinquency and a negative psychosocial outcome, it is important that other variables like intelligence, academic performance, education, and demographics are also considered as they combine to enhance the predictive power of problem behaviours in the the school based environment which place children and youth at-risk.
The social, economic, and mental health conditions currently experienced by Aboriginal children and youth place them at disadvantage to experience what Canadian society takes for granted. The effects of poverty, unemployment, crowded living conditions, higher rates of illiteracy, and lower levels of income typifies the ecological context of many Aboriginal communities (Hagey,
Larocque, & McBride, 1989a, 1989b; Muir, 1991; Norris, 1990; The Social
Trends Analysis Directorate, 1991). As well, elevated prevalence rates of alcohol and drug abuse, suicide, and family violence characterize the social and mental health of some of these communities (Health and Welfare Canada
Medical Services Branch, 1992; Kirmayer, 1994a, 1994b; Northern Medical
5 � .- .
- f 1-.....II1II -�" 445-- 7 :_.•_____ .. _... ,
Services Research and Development Committee, 1991; Report of the Statistical
Data Technical Working Group Medical Services Branch, 1991; Report of the
Steering Committee Medical Services Branch, 1991).
The purpose of the present study is to determine the predictors of problem
behaviours which place children and youth at-risk. The relationship between
problem behaviours and the demographic, academic, and cognitive variables
are explored. The variables are analyzed independently and in conjunction as
predictors of problem behaviours. As well, between group comparisons of
problem behaviours are made. Problem behaviours are defined as those which
encompass behavioural and emotional problems and include item descriptors
such as verbal aggression, swearing, in school referrals for discipline,
motivation and interest, stealing, lying, bullying, cheating, passivity and
withdrawal, overactivity and restlessness, nervous habits, daydreaming, lack of
concentration, few friends in class, unacceptable personal habits and hygiene,
does not follow directions, and self-centredness.
Problem behaviours are measured by the Teacher Rating Scale of the
Behavior Rating Profile-2nd Edition (BRP-2; Brown & Hammill, 1990). Academic
achievement based on teacher judgments is measured by the Academic
Performance Rating Scale (APRS; DuPaul, Rapport, & Perriello, 1990).
Estimates of the child's non-verbal reasoning ability and cognitive development
are assessed by the Matrix Analogies Test-Short Form (MAT-SF; Naglieri,
1985b), with the collection of demographic data pertaining to age, sex, and
school facilitated through a teacher-completed student identification
questionnaire. In conclusion, it is hoped the present research will facilitate the
development of policy initiatives and program planning which will enable at-risk
children and youth to experience success at school and in life.
6 CHAPTER TWO
LITERATURE REVIEW
In North America, over thirteen different cultural areas specific to Aboriginal peoples have been identified (Driver & MasseY,1957). Research corroborating the internal cultural similarities represented by these areas has been demonstrated ecologically and linguistically, lending support and corresponding to the cultural areas delineated. The thirteen generally agreed to
Aboriginal cultural areas which correspond to linguistic and ecological distinctiveness in North America and identified by McShane and Berry (1988) are:
1. Central and Eastern Arctic 2. Western Arctic 3. Yukon Sub-Arctic 4. MacKenzie Sub-Arctic 5. Eastern Sub-Arctic 6. Northwest Coast 7. Plateau 8. Plains 9. Prairies 10. East 11. California 12. Great Basin 13. Southwest (p. 388)
In discussing issues relevant to assessment in Aboriginal populations, as well as the mental health of these same populations, the literature will draw on the North American experience with reference to comparative studies both including and excluding indigenous populations. Reference to Canadian developments in research among Aboriginal people will refer to the cultural areas specific to Canada or common both to Canada and the United States
(Driver & Masey, 1957). An important distinction is made by this point: national borders are not the same as cultural boundaries and a Native person living in
Saskatchewan might have more in common culturally with a Native in South
7
���i"""'''''t1''/i� 1 � !r ,.� "; i
- � , ------�------Dakota than a non-Native living in Saskatchewan.
Simply stated, a member of the linguistic group represented by the language
Dakota, is part of the same Plains Indian cultural group as the Cree speaking individual in southern Saskatchewan and served by the same tribal council, a point alluded to in a statement provided by Gordon Lobe (cited in Schwean &
Saklofske, 1994), Superintendent of Education for the Saskatoon Tribal
Council, as part of the STC exceptionality study project background:
There are seven First Nations in the Saskatoon Tribal Council [John Smith, Kinistin, Mistawasis, Moosewoods, Muskeg Lake, One Arrow, and Yellow Quill], each of them representing a jurisdiction autonomous unto
itself.....The seven First Nations in this district include three different language groups (Cree, Saulteaux-Ojibway, and Dakota). The sizes of
the different communities vary from 200 to 800 total population .... (p.1)
Returning to the cultural and linguistic areas of Aboriginal populations in North
America, it is incumbent given the variance in cultural areas due to ecological and linguistic variables, that the effects of acculturation are recognized as influencing the development of indigenous cultures. Moreover, it is important to realize that acculturation has not affected all Aboriginal cultures in the same way and the degree or extent of contact has affected Aboriginal cultures differentially.
Acculturation, which refers to the replacement of traditional values and beliefs with those of the dominant culture, can occur in two differential contexts.
In one way, the culture of the Aboriginal group can become more like the
Western European norm and, in this ethnocentric sense, become closer to the norms of the larger more dominant society (McShane & Berry, 1988). In another, more debilitative way, acculturation can lead to the breakdown of the cultural and psychological functioning of a people (McShane & Berry), a topic which will be considered under the rubric of community mental and social health.
8 2.1 Demographics of Aboriginal People in Canada
As already has been stated, there are over nine different Aboriginal cultural groups in Canada. Within these nine different cultural groups exist over 606
Indian bands, with an average population of 550 people per band, that speak approximately 58 different dialects derived from 11 linguistic areas (Report of the Steering Committee Medical Services Branch, 1991; Health and Welfare
Canada Medical Services Branch; 1992). Furthermore, Canadian government legislation has delineated these cultural areas into four major Native groupings.
There are status Indians, which refers to those who are registered under the
Indian Act, of Canada; non-status Indians, meaning those who were not registered under the Indian Act of Canada; Metis, who are of both Indian and non-Indian ancestry; and the Inuit, who reside mainly in the Northwest
Territories and Northern Labrador. For the purposes of this study, it should be noted that the Saskatoon Tribal Council is comprised of status Indians who are living on the reserve and therefore, are registered according to the I ndian Act of
Canada.
2.1 .1 Population Distribution
Demographically, Aboriginal people are unique within Canada. At time of first contact with the Europeans, it is estimated the Aboriginal population numbered over 200,000. However, due to disease (tuberculosis) and war, this population base dwindled to about 80,000 in the late 1870's and then superseded the precontact population in the 1960's, to about 220,000 (Norris,
1990). Currently, it is estimated Aboriginal people number over 712,000, which is approximately 2.8% of the total Canadian population and of the total
Aboriginal population, only 28% or 159,665, live on-reserves. In Saskatchewan, over 53% of the Aboriginal population, by birth, reside on-reserves. When the
9 statistics compiled by Muir (1991) for the Department of Health and Welfare on
registered Indians are consulted, it is found that status Indians represented
1.7% of the total Canadian population in 1988, compared with 1.3% in 1981.
In Figure 2.1, the projected population trend for status Indian and Inuit
people in Canada according to Indian and Northern Affairs Canada (Hagey et
aI., 1989a) predicts a higher growth rate for both population groups. Due to Bill
C-31 the off-reserve status will , Indian population almost double its proportion of
the population from .41 % to .77%, with the higher growth rates in the status
Indian population between the years 1985 to 1990 primarily due to the effects of
Bill C-31 (Report of the Statistical Data Technical Working Group Medical
Services Branch, 1991). Thus not only will Aboriginal people's social and
political influence develop as their population grows, but because of the
increasing off-reserve population, federal and provincial governments may find
themselves having to provide services to off-reserve Indians which are similiar
to those provided to on-reserve Indians (Hagey et aI., 1989a, 1989b).
a Total Canadian II Status Indian Illill)lnuit
8 .
p 7
e 6 r
c 5
e 4 n �: �:::::�:��=:==:=��::�:::==�:-:::�i:==::: :::�:::� ::;���::.�-- :�:�::,:,-,:;::::: ::: ······· ··· ...... ····· ····· ·· .. ······ ·· v .. ·.. 1 "" ·)E )E ·v· ge )( � O+------�------�------�------�------1981 1986 1991 1996 2001 Year
Figure 2.1. 'Percentage Growth Rate of Status Indian, Inuit, and Total Canadian Populations. Source: Report of the Statistical Data Technical Working Group Medical Services Branch, 1991.
10
so In Figure 2.2, the distribution of Native people in Canada varies considerably with both the Northwest Territories and Yukon having the highest proportion of their populations claiming Native ancestry at 58% and 21 %, respectively (Norris, 1990; The Social Trends Analysis Directorate, 1991).
Provincially, Saskatchewan and Manitoba have the highest proportion of Native people at 8%, followed by British Columbia and Alberta at 5%, and central and eastern Canada between 1 % and 2% (Norris, 1990; The Social Trends
Analysis Directorate, 1991). When the population demographics are consulted specifically for Saskatchewan, remembering that Native people comprise 8% of the population, it is found that over two-thirds of the Native population for the province reside in Northern Saskatchewan, an area equivalent to Statistic
Canada's census division no. 18 (Northern Medical Services Research and.
Development Committee, 1991). In other words, 5 to 6% of Canada's Native population resides in northern Saskatchewan and 2 to 3% of Canada's Native population resides in southern Saskatchewan.
I Region I
Territories EPlIIIIII!IIIIlIIIIIIIIIlIIIIII!IIIIlIIII!III!lIIIIIIIIIiIIIIIIlIIIIIlIIIIISiIIIIIIlIIIIIlIIIIIIIIIIIIIIlIIIIII!IIII_IIIIIII!IIIIIIIIIIIIIIIIIIIIIIIIII!IIIIIIIIIIIIIIIIIII!IIIIIIII!II!I British Columbia Alberta
Saskatchewan . Manitoba Ontario Quebec Atlantic
0.00% 10.00% 20.00% 30.00% 40.00% 50.00% I a,On-Reserve 1II0ft-Reserve
Figure 2.2. Proportion of Persons of Native Origin and Total Canadian Population by Region. Source: The Social Trends Analysis Directorate, 1991.
11 In Figure 2.3, the number of Native people by region in Canada are compared. It is found that 42% of the total Native population reside in Ontario,
36.3% in British Columbia, 36.1 % in Saskatchewan, 27.6% in Manitoba, 27.2% in Alberta, 24.3% in Quebec, 9.8% in Atlantic Canada, and 6.7% in the
Territories (The Social Trends Analysis Directorate, 1991). Furthermore, the majority of Native people who live in Saskatchewan reside off-reserve as compared with those living on-reserve.
Territories __---
British Columbia : :.: :.,;,;';',., ,« , .. Alberta .
Ontario ...... •......
.. :: : Quebec :,.. .:;:: ...... :,.;,.: Atlantic : , 0.00% 5.00% 10.00% 15.00% 20.00% 25.00% 30.00% 35.00%
a On-Reserve II Off-Reserve [ill] % of Total Canadian Pop.
Figure 2.3. On-Reserve, Off-Reserve, and Total Canadian Populations by Region. Source: The Social Trends Analysis Directorate, 1991.
2.1.2 Birth Rates
It is generally agreed the determination of population growth through birth rates is somewhat fraught with difficulties, owing to definitional criteria, reporting, and data collection methods. Nevertheless, it is still an effective procedure for the charting of population growth, and at the moment, is the only technique available. Earlier it was mentioned that after contact with the
Europeans, the Aboriginal population in Canada dropped off significantly.
However, between the 1940's and 1980's, the Aboriginal population growth in
12
>",_ .�, I ',',_'.�' •.• l�,...",
--,,,,...- . .., { ,------��� Canada is estimated to have reached 205%, compared with 109% for the
Canadian population as a whole (Norris, 1990). In 1988, the crude birth rates
for the registered Indian population, which takes into account all births
irrespective of whether the birth was live or not, was 29.3 per 1,000, compared
with the Canadian crude birth rate of 14.5 per 1,000 (Muir, 1991). Similarly, the
natural increase, which is the excess of live births over deaths per 1000,
according to Muir has been increasing for the registered Indian population.
In the Canadian population, the average mean age is 34, whereas for
Canada's Aboriginal population the average mean age is 24. Similarly, only
47% of the Canadian population is under the age of 30, whereas over 68% of
the Canadian Aboriginal population is under the age of 30 (Norris, 1990).
Subsequently, the effect of the higher natural increase in birth rates has
resulted in a younger age structure among Canada's Status Indian population
in comparison to the total Canadian population, see Figure 2.4.
70.00% a Status Indians: 0-1 4 years
II Total Canadian: 0-14 years p 60.00% [ill) Status Indians: 1 5-64 years ; 50.00% [ill Total Canadian: 1 5-64 years c 40.00% Status e IIlIIIII Indians: 65+ years
n 30.00% a Total Canadian: 65+ years t a 20.00% 9 e 10.00%
0.00% 1981 2001 Year
Figure 2.4. Age Structure of Status Indian and Total Canadian Populations. Source: Report of the Statistical Data Technical Working Group Medical Services Branch, 1991.
13
, The implication of the younger age structure in the Aboriginal population for the education system is best understood in relation to the increasing Aboriginal proportion of the population (Hagey et al., 1989a):
There will be a dramatic increase in the number of Indians on-reserve in the education system not only through higher retention rates but due to a continuous increase in the school age population (5-21), with 21,000 more in this age group in the year 2001 than in 1991. (p. 17)
Because of this, there is a greater demand for local band run schools to be developed from community education resources. Therefore, it is imperative that educational services be adequately developed and implemented so as to serve the needs of the on-reserve school age population.
Even though both the crude and natural increase in birth rates is higher for
Aboriginal people, as compared with the Canadian population, trends in fe�ility rates indicate that like the Canadian levels, Aboriginal fertility levels are declining (Norris, 1990). Part of the reason why this is so is because of the increased use of birth control, considerations relating to the timing, size of families, and marital instability (Norris). Likewise, if the total fertility rates are consulted, it is found the fertility rate of both Indian and Inuit people declined more rapidly than the Canadian norm and are lower in urban areas than rural
(Norris). Nevertheless, even though total fertility rates are declining, Aboriginal families, especially among the Inuit, tend to have more children compared with the Canadian population:
The comparison of total fertility rates indicates that Inuit fertility has been and remains higher than that of registered Indians. Data from the 1981 Census on "children born to ever-married women" clearly confirm this relation. In 1981 the average number of children born to ever-married women, was highest for Inuit, at 4.7, followed by 4.1 for status Indians, 3.6 for Metis, 3.0 for non-status Indians, and 3.9 for non-status and status Indians combined. The average number of children for Canadians in general was much lower at 2.5, compared to 3.9 for Native people as a whole. (Norris, p. 45)
14 When data are consulted regarding family size, it has been found that status
Indian families tend to be larger than the national average, on-reserve families
are than those families to Bill the of larger off-reserve, and due C-31 , number
Indian families will increase as the year 2001 approaches (Hagey et aI., 1989a).
2.1.3 Dependency Ratio
One result of the age composition outlined is an increase in the dependency
ratio among Aboriginal people. The dependency ratio is calculated by
determining the total population who has little or no wage earning potential, in
other words, the dependent, and dividing it by the total population which has
wage earning potential, the independent, using the following formula:
[Total Population (0-14) + (65+)ITotal Population (15-64)] * 100, (Graham,
1987). Thus given the dependency ratio calculation outlined, if a dependency
ratio exceeds 100, then there exists more than one dependent for each adult of
working age. However, it is important to note that wage earning potential does
not imply the adult of working age has to be gainfully employed.
At present, the Canadian dependency ratio is 46 per 100, whereas the
dependency ratio for on-reserve Indians is 76 per 100 and for off-reserve
Indians, 62 per 100. In Saskatchewan, which has the highest dependency ratio
of all provinces and territories, the dependency ratio for on-reserve Indians is 91
per 100 and for off-reserve Indians, it is 81 per 100. The significance of this
statistic, according to The Social Trends Analysis Directorate (1991) is that:
Because of this age structure, the dependency ratios of persons of Aboriginal origin are significantly more youth-biased than the general population. As a result, as the current youth generation of Aboriginal Canadians ages and enters the labour market, employment and job expectations will likely become an even more critical policy issue. (p. 6)
The implication for educational services is that these services will need to -be
matched with the current demands dictated by market and societal pressures.
15
I · ...... -----� _" - _._.".__._ �
.. • •...•• • • 0, , • , ,"' • •
2.1.4 Living Conditions
In the past, many on-reserve dwellings for Indian people had inadequate sewage systems and a higher proportion of homes without central heating and running water compared with the national average (Report of the Steering
Committee Medical Services Branch, 1991). When considering the quality of living conditions, another important indicator is the number of individuals who reside within a crowded dwelling. Accordingly, "a crowded dwelling is defined as any dwelling occupied by more than one person per room" (Hagey et aI.,
1989a, p. 13). Although the percentage of crowded dwellings has been decreasing for the total Aboriginal population, the rates of crowded dwellings among Aboriginal people are significantly higher than the national average
(Hagey et al,). Thus 28.9% of on-reserve Indians live in crowded dwellings, compared to 11.3% for off-reserve Indians, and 1.8% for the total Canadian population, see Figure 2.5. What these percentages demonstrate is that on reserve status Indians live in crowded dwellings at 16 times the Canadian rate.
C
. r 50.00% a Total Canadian o On-Reserve w 111 d 40.00% (illill Off-Reserve e Dlnuit d 30.00%
D w 20.00% e I I 10.00%
1981 1986
Figure 2.5. Crowded Dwellings: On-Reserve, Off-Reserve, Inuit, and Total Canadian Populations. Source: Report of the Statistical Data Technical Working Group Medical Services Branch, 1991.
16 As well, there are more single parent families of Native origin living in
Canada today than the national average, see Figure 2.6. Approximately 19% of
all Native families in Canada are single parent households, compared with the
national average of 12.7% in 1986. Of the 26.9% status Indian families that are
of single parentage, over 86% are single mothers, with 60% of these women
living off-reserve (Hagey et aI., 1989a).
30.00% a Male • Female Ii) Total 20.00% .
0.00% Total Canadian On-Reserve Off-Reserve Inuit
Figure 2.6. Single Parent Families: On-Reserve, Off-Reserve, Inuit, and Total Canadian Populations. Source: Report of the Statistical Data Technical Working Group Medical Services Branch, 1991.
Another indicator of social need and living conditions is the number of
children who are in the care of child welfare authorities. Even though the
number of Native children in the hands of child welfare agencies is decreasinq,
the proportion of children in care is already higher than the average, putting
extra demands on limited resources and generally indicating overall poorer
living conditions, see Figure 2.7. With higher rates of single parent families,
crowded dwellings, and children in care, there may be a tendency for an overall
greater demand to be put on social service resources.
17
.. _ I ----��--- _ .. 6 a On-Reserve
% II Total Canadian 5
o f 4
C h 3 i I 2 d
r
e 1
n
o 1981198219831984198519861987 Year
Figure 2.7. Children in Care for On-Reserve and Total Canadian Populations: 1981-1987. Source: Report of the Statistical Data Technical Working Group Medical Services Branch, 1991.
2.1.5 Education
When examining the education level of the total population aged 15 years
and over, it is found that approximately 26% of all on-reserve and off-reserve
Indian people have less than grade nine level of completion, compared with
17.3% for the Canadian population, see Figure 2.8. The education level is also
lower among those who live on-reserve, with over 43% of those surveyed
reporting having attained grade nine or lower. This figure is even higher in
Saskatchewan where 50.6% of all Native people living on-reserve have a
grade nine or less, which is the second lowest education level in the country,
with only Manitoba reporting a rate of 52.3% for their on-reserve population
(The Social Trends Analysis Directorate, 1991). When the same measure is
used for status Indians, it is discovered that 37% of all status Indians have less
than grade nine, which is twice the proportion reported for aU of Canada (Report
of the Steering Committee Medical Services Branch, 1991). As well, when the
18
.t: 4''''Ii,''I �:� l __ __ criteria of having received a high school diploma is used to determine
education level, only 8.1 % of the total Native population report having attained
a secondary diploma, compared with the Canadian total of 12.8%. In
Saskatchewan, only 1.6% of the on-reserve population and 4.6% of the off
reserve population have attained a high school diploma, see Figure 2.9 (The
Social Trends Analysis Directorate, 1991).
As well, it has been noted that functional illiteracy is declining in Aboriginal communities (Report of the Steering Committee Medical Services Branch,
1991), with The Social Trends Analysis Directorate (1991) disclosing: "the
educational characteristics of persons of Aboriginal descent living off-reserve
appear to be closer to those of the Canadian population as a whole than those
of Aboriginal people living on-reserve" (p.10).
I Region I
British Columbia
Alberta '::::':.:':::.:':::.::".:":::.":: ...
Manitoba
. Ontario ::;.; ;:.: ::.: ;:.::::
.... . Atlantic
I 0.00% 10.00% 20.00% 30.00% 40.00% 50.00% 60.00%
a On-Reserve II Off-Reserve Ilili] Total Canadian Population
Figure 2.8. On-Reserve, Off-Reserve, and Total Canadian Populations With Less Than Grade 9. Source:The Social Trends Analysis Directorate, 1991.
19
..... '\ I ---- �� ,")
. .v« ------..:�- ..,__.. 1, -
-
-
-
I Region I
0.00% 10.00% 20.00%
a On-Reserve III Off-Reserve iii Total Canadian Population
Figure 2.9. On-Reserve, Off-Reserve, and Total Canadian Populations with High School Diploma. Source: The Social Trends Analysis Directorate, 1991.
2.1.6 Economic Status
Over 19% of Aboriginal people in Canada have no income compared to
18% for the off-reserve population, 24% for the on-reserve population and 13% for the total Canadian population, see Figure 2.10 (The Social Trends. Analysis
Directorate, 1991). Similarly, 51 % of all off-reserve Aboriginal people have an
income of less than $10,000; with 68% of the Aboriginal population living on
reserve having an income of less than $10,000, compared with 39% for the total
Canadian population. Regionally, Saskatchewan has one of the highest
proportions of Aboriginal people living off-reserve with income less than
$10,000, at approximately 60%; the highest rate of families living off-reserve
below a combined income of $25,000 at 64%; and the second highest rate for
on-reserve families living on a combined income of less than $25,000 at 75.8%.
Manitoba's on-reserve rate was only slightly higher at 76.7%, with the total
20 ' ......
Canadian proportion of families combined income below $25,000 being 31% or
one in three (The Social Trends Analysis Directorate). What these statistics
describe is a population which has a higher proportion of people with either no
income or a lower income than the national average, and of those that do have
a higher income, only 12.6% report having an income greater than $20,000,
compared with 31.1 % for the Canadian population having this income level
(Muir, 1991; The Social Trends Analysis Directorate).
Off-Reserve
On-Reserve
Total Canadian
0.00% 15.00% 30.00% 45.00% 60.00% 75.00% % With Family Income
DNolncome II Less than $10,000 Iillill Less than $25,000
Figure 2.10. Income Level of Families On-Reserve, Off-Reserve, and Total Canadian Population. Source: The Social Trends Analysis Directorate, 1991.
2.1.7 Unemployment and Labour Force Participation Rates
The definition offered by Statistics Canada for the concepts labour force
participation and unemployment were not developed explicitly for Native
populations. Because of this, activities such has hunting and trapping, which
are traditional pursuits of Native people, are not considered labour force
participation if the products of such endeavours are either bartered or
consumed and therefore, do not enterthe market economy (The Social Trends
Analysis Directorate, 1991). On account of this, the validity of labour force and
21 I
...)
unemployment statistics might be circumspect. Nevertheless, higher rates of
unemployment exist among Canada's Native population, see Figure 2.11.
It should also be noted that the on-reserve Aboriginal population in 1986
had a 34.9% unemployment rate compared with the off-reserve rate of 20.5%
and a Canadian rate of 10.3% with Saskatchewan's on-reserve and off-reserve
populations posting unemployment rates of 33.6% and 27.9%, respectively, see
Figure 2.11 (The Social Trends Analysis Directorate, 1991). In terms of
participation rate in the labour force, it was found the on-reserve and off-reserve
labour force participation rates were 43.2% and 65.1 %, compared with the
Canadian rate of 66.5%, with Saskatchewan's on and off-reserve rates being
36.7% and 56.5%, respectively. What these statistics describe for Canada's
Aboriginal off-reserve population is an unemployment rate which is twice as
high as the Canadian rate and a labour force participation rate which
approximates the Canadian norm.
I I Region I
0.00% 10.00% 20.00% 30.00% 40.00% 50.00%
a On-Reserve II Off-Reserve Iillilll Canadian Population
Figure 2.11. Unemployment Rate of On-Reserve, Off-Reserve, and Total Canadian Populations. Source: The Social Trends Analysis Directorate, 1991.
- _;:;:�':':.�:"V ;,��:;�m��.l!�,�',. , 22 .J
For on-reserve populations, the unemployment rate is over three times
higher than the Canadian rate, see Figure 2.12. As well, the labour market
participation rate is over 20% lower compared with the Canadian national rate.
However, it should be realized that the validity of the data as previously
described is limited, and if hunting and trapping outside of the market economy
is considered, especially among on-reserve dwellers where hunting and
trapping is more of a lifestyle choice and occurs as part of daily life, it is
suspected some of the rates reported would change.
I Group I
Inuit
On-Reserve
Total Canadian
0.00% 15.00% 30.00% 45.00% 60.00% 75.00%
a Employed Ell Unemployed IilliliI Not in Labour Force
Figure 2.12. Labour Force Activity for On-Reserve, Off-Reserve, lnuit, and Canadian Populations. Source: Report of the Statistical Data Technical Working Group Medical Services Branch, 1991.
Regardless of the possible validity error in the statistics, as the increasing
age-youth structure changes, remembering the Canadian trend is towards a
greater percentage of elderly people in the population, higher rates of
unemployment and dissatisfaction could result if resources are not allocated to
serve the needs of Aboriginal youth (Hagey et aI., 1989b):
As the Indian population ages and young people enter the work force, heavy demands will be placed on Indian communities to provide suitable
23 employment opportunities. By the year 2001, there will be 56,000 youth on-reserve of working age (17-24), 5,000 more than in 1991. Finding work for Indian youth will become an increasing challenge. Failure will add to the social assistance roles (p. 18)
Furthermore, the impact of even higher unemployment rates on the mental and
social health of Aboriginal communities would be debilitative and therefore, it is
imperative that government agencies and Indian communities work together to
solve the current maladies facing Canada's Aboriginal youth.
2.1.8 Language
The final topic to be covered in an examination of Aboriginal social and
economic demographics is language. Of those Aboriginal people living off
reserve, only 11 % report an Aboriginal language as their mother tongue,
compared with 45% for the on-reserve population (The Social Trends Analysis
Directorate, 1991). When the data are consulted for those persons of Aboriginal
origin who speak an Aboriginal language at home, it is found 7% of the off
reserve population use their mother tongue as the primary language in the
home compared with 35% for on-reserve dwellers. These figures seem to
indicate that those who live on-reserve both retain and use their Aboriginal
language more frequently than those who are living off-reserve.
2.1.9 Summary
Because of the previous demographic trends outlined and the unparalleled
position in which Aboriginal people find themselves, there is good argument for
considering these demographic indicators when attributing and delineating
casual factors to conditions of mental health in Aboriginal communities. It is
found that Aboriginal people are distinct from the broader context of Canadian
society in many ways. The Aboriginal population in comparison to the total
Canadian population has: a higher population growth confirmed by birth and
24
I - � J
fertility rates that exceed the national rates; a higher proportion of younger
people with a younger age structure by 10 years; lower overall provincial and
territorial education levels; and a dependency ratio which is even more youth
biased than the national average indicating an even greater demand among
youth for scarce labour market resources, where current unemployment rates
exist between 20 to 30%. Moreover, when these same statistics are consulted
and the conditions facing on-reserve and off-reserve Indians are compared, it is
discovered that on-reserve Indians are worse off than off-reserve Indians and
the Canadian population as a whole.
But the social and economic demographics described do not occur in
isolation. Rather they describe a society or culture in which children, youth, and
families are at risk for a negative psychosocial developmental outcome. To.
understand the context which characterizes mental health in Aboriginal
communities, it is necessary to understand the impact of social pathologies, like
morbidity and mortality, on these same communities (Report of the Statistical
Data Technical Working Group Medical Services Branch, 1991). By so doing,
the mental health, and the predisposing at-risk factors, will be better understood
in light of the socio-economic conditions currently facing Aboriginal people.
2.2 Mental Health in Aboriginal Communities
The rapid effects of acculturation have significantly altered the ecological
context of Aboriginal people. Poverty is common in many communities and its
causes are the social inequalities and discrimination which confront Aboriginal
people today (Northern Medical Services Research and Development
Committee, 1991). Higher rates of unemployment and illiteracy coupled with
poorer living conditions are characteristic of the ecological context encountered
in many Aboriginal communities. The result of this rapidly shifting ecological
25
, context has seen the increased prevalence of mental and social health problems among numerous Aboriginal communities in Canada. But it is important to remember that culture change, in itself, is not reponsible for all aspects of psychopathology that any particular society or community might experience. As Kirmayer (1994b) remarks:
All communities create and control certain forms of deviance to define and integrate themselves. Yet, social pressures on the individual labelled as deviant may be so powerful that they have deleterious effects on their mental health and create pathology. (p. 45)
Therefore, in order to understand mental health in Aboriginal communities, it will be necessary to explore a concept of mental health which is sensitive to the
Aboriginal perspective.
2.2.1 Definitional Criteria
In 1990-91, a steering committee on Aboriginal mental health was initiated.
This group consisted of representatives from Indian and Northern Health, the
Department of Indian and Northern Development (OlAND), the Assembly of First
Nations Health Commission, the Labrador Inuit Health Commission, the
Nishnawbe Aski-Nation, River Desert, the Manitoba Assembly of Chiefs, the
Meadow Lake Tribal Council, the Alberta Indian Health Care Commission, the
Sto:Lo Nation, and the B.C.Congress of First Nations (Report of the Steering
Committee Medical Services Branch, 1991). The mandate of this committee, formed under the auspices of Indian and Northern Health Services, Medical
Services Branch, was to document current concerns and issues among the
Aboriginal community surrounding mental and social health. Because of this, an attempt was made to provide direction to the federal government response so as to reflect the goals and aspirations of the communities it serves (Report of the
Steering Committee Medical Services Branch). As a result, the Report of the
26 Steering Committee Medical Services Branch developed the following
definition of mental health which depicts, in part, the Aboriginal concept of
mental health:
Among the First Nations and Inuit communities the term mental health is used in a broad sense of describing behaviours which make for harmonious and cohesive community and the relative absence of multiple problem behaviours in the community, such as family violence, substance abuse, juvenile delinquency, and self-destructive behaviour. It is more than the absence of illness, disease or dysfunction-it is the presence of a holistic, psychological wellness which is part of the full circle of the mind, body, emotion and spirit, with respect for tradition, culture and language. This gives rise to creativity, imagination and growth, and enhances the capacity of the community, family group or individual to interact harmoniously and respond to illness and other adversity in healing ways that resolve conflicts constructively, promote improved function and the healthy development of children. (p. 6)
With this definition of mental health, it is incumbent that the multiple problem
behaviours surrounding social and metal health dysfunction be explored in
relation to First Nation communities.
2.2.2 Indicators of Mental Health
Common indicators of mental health in a community are deaths, suicide,
substance abuse, family violence, and childhood/adolescent problems. It is
known that low levels of employment and recreation activities can intensify the
chronicity and severity of violence and substance abuse among teenagers who
perceive life as meaningless and without purpose (Report of the Steering
Committee Medical Services Branch, 1991). Because of the possible , relationship that exists between socio-economic demographics and mental
health indicators, it is necessary, to understand the complexity of the issues
facing Aboriginal communities, to examine more fully the indicators of mental
health.
27
-, I 2.2.2.1 Mortality
The current mortality rate for Aboriginal populations is much higher than the
Canadian average. Although both Indian and Inuit life expectancies have risen
over the years, they still remain substantially lower than the national average. At
the turn of the century, the average life expectancy for registered Indians of both
sexes was only 33 years, later in 1982 to 1985, it was estimated that for the
same population, life expectancy had risen to 64 and 72.8 years for males and
females, respectively (NorriS, 1990). It is expected by the year 2001 the average
life expectancy of Native Canadians will be 6 years or less than the national
average, at 70 versus 76 for men, and 77 versus 83 for women (Hagey et aI.,
1989a). In comparison, the Canadian life expectancy, calculated during the
period from 1984 to 1986, was 73 and 79.8 years for males and females,
respectively. That there should be such a rapid increase in the life expectancy
of registered Indians is partially due to lower infant mortality rates. Although
infant mortality has been decreasing among Aboriginal populations, it is still
higher than the Canadian proportion of the population. As a result, the
registered Indian mortality rate is 2.2 times higher than the Canadian rate and
for the Inuit, it is 2.6 times the Canadian rate (Muir, 1991). In Saskatchewan, the
infant mortality rate has been declining from a high of 45.8 per 1,000 live births
to a rate of 15.4 per 1,000 live births in 1986 (Health and Welfare Canada
Medical Services Branch Saskatchewan Region, 1989).
The significance of infant mortality is that it is closely linked to socio
economic status. Because poverty affects one's health, mothers who are
afflicted by poverty can have premature and/or babies of lower birth weight due
to poor nutrition resulting in higher rates of infant mortality (Muir, 1991). Even
though infant mortality rates are decreasing for Canada's Aboriginal population,
although still higher than the national rate, the mortality among older ages is
28
� I either staying the same or increasing in some age groups. As to the importance
of these figures, Norris (1990) explains:
This pattern of relatively little decline in mortality at older ages, compared to the great strides made in infant survival, reflects that the major causes of death for registered Indians are not associated with disease, but rather with "accidents, poisoning and violence," and many of these deaths are alcohol related. (p. 50)
In fact, between 1986 to 1988, the leading cause of death among the
registered Indian population was injury and poisoning at 31.2%, varying
between 49.4% for the Yukon and 20.5 % for the Atlantic region, with
Saskatchewan posting a rate of 36.8% (Muir, 1991). Moreover, when the cause
of death statistics are consulted for the total Indian population, injury and
poisoning accounts for 86% in the 15 to 24 year age group and 59% in the 25 to
44 year age group and 60% in the 45 to 64 year age grouping (Health and
Welfare Canada Medical Services Branch, 1992). In comparison, the leading
cause of death in the Canadian population for the same time period was
diseases of the circulatory system at 43%, with injury and poisoning ranking
fourth at 7.5% (Muir).
Moreover, if age-specific death rates are consulted, it is found that 64.2% of
the deaths among children between 1 to 14 years of age, and 87% of the
deaths among youth and young adults between 15 to 24 years of age are the
result of injury and poisoning, with violence and accidents resulting in death for
Indian men under 70 years of age occurring four to five times more frequently
than the Canadian average (Report of the Steering Committee Medical
Services Branch, 1991).
When suicide rates are examined, it is found that in Northern Saskatchewan
which is comprised of a population that is over two-thirds Native, the rate of
suicide is four times the provincial rate (Northern Medical Services Research
and Development Committee, 1991). In Canada, the rate of suicide among
29
�\ I Indian people, referring to on-reserve status and registered Indians, is five to six
times higher than the Canadian rate (Report of the Steering Committee Medical
Services Branch, 1991; Kirmayer, 1994b). Moreover, such a death is likely to
involve alcohol, be carried out with either a gun or by hanging, and to occur in
clusters among Native adolescents (Kirmayer). As to the significance of these figures, regardless of whether one views suicide as a result of individual or familial psychopathology, or viewed from a sociological perspective as the
result of political disempowerment, acculturative stress, or low socioeconomic
status (Kirmayer), it is a powerful indicator of the mental well being of a
community (Report of the Steering Committee Medical Services Branch).
2.2.2.2 Alcohol and substance abuse
When the incidence rates of alcohol related offences for Aboriginal people
are examined, they are found to be five times the national average (Northern
Medical Services Research and Development Committee, 1991). As well, over
80% of the diagnosed psychopathologies are of addicted people. Moreover, when one compares non-Native and Native homicides, the vast majority of
Native homicides are alcohol related showing no evidence of psychosis or
sexual deviancy, which are both common among the non-Native population in
Canadian penitentiaries (Proceedings of the Canadian Psychiatric
Association's Annual Meeting, 1982). In terms of child psychopathology, it is
informally estimated that in Aboriginal communities over 25% of the births are
affected with fetal alcohol syndrome, the leading cause of mental retardation
today (Report of the Steering Committee Medical Services Branch, 1991).
30 2.2.2.3 Family violence and child sexual abuse
Society has seen a growing trend to uncover the issues pertaining to family violence and child sexual abuse. Aboriginal people have participated in this process by coming to terms with the abuse experienced in the residential schooling system, religious institutions, and other situations conducive to the promotion of child abuse (Kirmayer, 1994a). Issues surrounding family violence and child abuse continue to be of major concern to many Aboriginal communities (Report of Steering Committee Medical Services Branch, 1991).
As to what factors place a child at risk for abuse, Kirmayer states:
The majority of sexual abuse is initiated when children are less than seven years old; 70% of abuse takes place more than once. Many offenders engage in more than one type of sexual deviancy. For example, up to 50% of fathers referred to a clinic for incest also abused children outside the home. At present, it is not clear what factors increase a child's risk except female gender, separation from mother and poor relationship with mother (these last two may be consequences rather than risk factors). (p. 32)
At present, it is generally conceded the punitive and adversarial nature of the justice system does not deal adequately with cases of abuse and is tearing some communities apart (Report of the Steering Committee Medical Services
Branch). Furthermore, the impact of poverty, isolation, and demoralization, coupled with alcohol and substance abuse, appear to contribute to family violence and child abuse in many Aboriginal communities (Kirmayer).
2.2.2.4 Summary
Given the current situation facing the mental health of Aboriginal communities, it is clear that social, economic, and cultural factors do not occur in isolation, but rather impact in conjunction on Aboriginal communities. Because of these conditions, one would necessarily expect that all traditional values and beliefs would have been subverted by Euro-American or Euro-Canadian
31 culture. Although in many situations the effects of geographic relocation and
residential schooling has resulted in the break down of the extended family and
traditional values, beliefs, and practises (Report of the Steering Committee
Medical Services Branch, 1991), not all cultural processes and structures have
been abrogated. Rather, individual coping styles (LaFromboise, 1988;
McShane, 1987; McShane and Berry, 1988) and other traditional cultural
structures and processes remain, which cause some Aboriginal children and
youth to be resilient to disaffected emotional and behavioural engagement
thereby, not placing them at risk for failure in life or at school . Thus according to
McShane and Berry:
The remaking of the environment contrasts with the enduring of certain cultural structures and processes. These include the maintenance of some aspects of traditional structures of adult-adult and adult-child interactions, and some continuity of traditional forms of relations between mother and child. (p. 404)
If the argument posited by McShane and Berry (1988) is accepted, it must be
recognized that Aboriginal culture remains distinct from the North American
norm due to the retention of traditional cultural processes and structures. The
extent to which these processes and structures play themselves out are
questions that researchers, educators, psychologists, and health care
professionals must ask themselves when assessing and treating Aboriginal
populations. Similarly, historical, economic, and social factors have come to
bear on Aboriginal communities such that these environmental or external
factors have profoundly affected mental health in Aboriginal communities
(Report of the Steering Committee Medical Services Branch, 1991). The Steering Committee for the Medical Services Branch in their 1991
report concluded Native people were no more prone to diagnosable psychiatric
disorders even though there exists a disproportionate number of Native people
diagnosed as schizophrenic. Although the effects of alcohol and substance
32
, .1 abuse were fairly wide ranging, the Steering Committee concluded the uses of such substances concealed deeper social psychopathologies and where alcohol had been controlled, other problem behaviours previously not present, began to emerge (Report of the Steering Committee Medical Services Branch,
1991).
It was also noted there exists higher rates of depression and failure to thrive among children. Even though indicators of mental health predict higher rates of psychopathology among children, under identification among behaviourally disordered youth continues to occur (Schwean & Greenough-Olsen, 1991). It is not surprising then that Aboriginal communities in Canada are demanding better services which translate into a more adequate share of resources as based on need.
It is important to remember the statistics for mental and social health in
Aboriginal communities are based on overall prevalence rates and not all
Aboriginal communities have a high prevalence of emotional and behavioural problems demanding immediate care. Given this situation, it is clear some
Aboriginal communities are worse off than others, if prevalence indicators for mental and social health are articulated on a community by community basis. It is implicit then that culturally fair assessment practices be adopted so as to accurately identify the needs and subsequent resource allocation within
Aboriginal communities, schools, and health facilities.
2.2.3 Epidemiological Studies
The most comprehensive epidemiological study of childhood and adolescent disorders in Canada was the Ontario Child Health Study (OCHS) conducted in 1983 in the province of Ontario. The purpose of the OCHS was to survey and estimate the prevalence of the "four childhood psychiatric disorders:
33 conduct disorder, hyperactivity, and emotional disorders (neurosis) in children,
4-16; and somatization in adolescents 12-16 years of age" (Offord et aI., 1991,
p. 31). The OCHS is found wanting in its applicability to Aboriginal child and
youth populations in Canada, however, because Aboriginal children and
adolescents living on-reserve were excluded from the sample ( Boyle et aI.,
1987). Since Aboriginal children were not included in the subject sample the
generalizability and the ecological validity of OCHS findings as applied to
Aboriginal groupings is questionable. The strength of this point is made
apparent when it is recognized the authors of the OCHS did not intend their
study to be representative of Aboriginal populations due to the lack of
instrument standardization (Armstrong, 1993).
Because of this, prevalence rates for specific disorders might vary across
cultures as a factor of psychopathology. Nevertheless, the OCHS provides
valuable information about the correlates of childhood psychiatric disorders
which can be used for comparative purposes with Aboriginal populations. To
address the shortcomings of the OCHS epidemiological survey, it will be
necessary to draw on research conducted in the United States on the American
Aboriginal youth population. Further research in Canada should focus on the
deficiency in the epidemiological literature and provide a more comprehensive
picture of child and adolescent psychopathology in Canada's Aboriginal
population.
2.2.3.1 Ontario Child Health Study
The OCHS was commissioned by the Province of Ontario's Ministry of
Community and Social Services (MCSS) so as to better plan and design the
dispensation of services (Boyle et aI., 1987). Having recognized the
purposefulness of such a study and the need for appropriate resource
34
� � I I j-
allocation, the MCSS according to Boyle et al.:
... wished to obtain estimates of the prevalence of emotional and behavioural disorders among children 4 to 16 years and information on the use of mental health, social, correctional, special education services, and medical care by Ontario's children. (p. 826)
Even though the ecological validity of the OCHS for the Aboriginal population is
circumspect, it is important the findings be considered not only for their
concordance but also for their possible variation, which in part could be
explained by socio-economic demographics, social and mental heath
lndlcatorscutture! differences, and assessment practises. Given these
considerations, it would be expected that research in Aboriginal child and
adolescent psychopathology should yield findings which are discrepant from
the total population. Moreover, the causal factors delineated would need to be
sensitive to the characteristics of the Aboriginal grouping under study, as it is
known that variations exist economically, socially, and culturally among the
various Aboriginal groupings.
2.2.3.2 Ontario Child Health Study: Findings and limitations
According to Boyle et al. (1987), a number otnmltanons apply to the OCHS
methodology. The use of clinical diagnosis for the confirmation of a disorder is
problematic since diagnosis was not standardized and left to the vagrancies of
the clinicians, thus allowing for differences in information gathering and criteria
invoked in the assessment process; secondly, emotional disorders were
derived through the collapsing of anxiety and depressive disorders into one
category as was done in the Isle of Wright epidemiological study by Rutter
(1989) making the measurement of child and adolescent depressive and
anxiety disorders not possible; thirdly, even though multiple perspectives were
used in the assessment process (Le. teacher, student, and parent ratings), little
35
"",-\ _jl
is known on how to integrate the data collected. On the other hand, the
strengths of the OCHS were the very high participation rates at 91.1 %, and low
refusal rates at 3.9% (Boyle et al.), which due to the low attrition, supports the
internal validity of the study.
The four psychiatric disorders assessed in the OCHS (Boyle et al., 1987)
were determined through checklists which were converted to a binary scale.
The binary scale then reflected either the presence or absence of a disorder. An
individual was then classified as disordered based on a positive rating by one
or both of the informants, teacher or parent in the 4 to 11 year age group and
youth or parent in the 12 to 16 year old age group.
The correlates of the disorders examined by Offord et at (1989, 1991) were
low income, rural/urban residency, family dysfunction, parent treated for nerves,
parent arrested, child's age, child's sex, whether the child repeated a grade,
and chronic medical iIIness.The correlates defined by Offord et al., (1989) are
as follows:
Low income:Total family income in the preceding year (1982) was <$10,000. Urban/rural: Urban areas are those with a population of >25,000. Rural areas include both small urban areas (population 3,000 to 25,000) and rural areas (population <3,000). Family dysfunction: A score above 26, on a range of 12 to 48, on the 12-
item General Functioning subscale .... Parent treated for nerves: Parent reports that self or partner was treated at some time for nerves or a nervous condition. Parent arrested: Parent reports that self or partner was at some time arrested or charged with an offence other than a traffic violation. Failed a grade: Parent reports that child failed or repeated a grade at some time during his or her school career. Chronic medical illness: Parent reports that child has one or more illness or conditions which are usually chronic in duration (>6 months). (p. 196)
The major finding of the analyses was the difference informants made in
determ ining the strength of the relationship between the correlates and a
specific disorder. In this respect, it should be noted that parents completed
36
, questionnaires for both the 4 to 11 and 12 to 16 year age groups, teachers completed questionnaires only for the 4 to 11 year age group, and due to the unreliability of administering self-report questionnaires to the 4 to 11 year age group, only youth completed questionnaires for the 12 to 16 year age group.
Given the questionnaires completed by the informants for the age groups specified, it was found that child self-reports, youth self-reports, parent-reports, and teacher-reports varied significantly in the identification of a disorder as a function of the informant (Offord et aI., 1989). For instance, in the 4 to 11 age group, teacher- and parent-identified conduct disorder and hyperactivity varied differentially with teachers evidencing higher prevalence rates in their identification. There was, however, agreement between teacher and parent ratings as to the elevated prevalence of both externalizing disorders in the male population sample. When internalizing emotional disorders were considered, it was found that parents identified the females more often than the males, the reverse of the teacher identification rates (Offord et aI., 1989).
Again there were differential rates of identification in the 12 to 16 year age group. In this group, a youth self-report form and parent checklist was completed. It was found the youth-identified higher prevalence rates of conduct disorder among males and higher prevalence rates of emotional disorders and somatization among females than parent-identified cases. Both parents and youth-identified conduct disorder as being more common in males, with emotional disorders and somatization more common in females for the same age group (Offord et al., 1989).
When the patterns of correlates were examined, it was also found to vary among informants and the disorder identified. For example, the relationship between low income, parent arrested, and age 12-16 years and parent identified conduct disorder was significant (Offord et al., 1989). Conversely, a
37 ! ! ...... _...... _?liiiiIItlll4'•• i .. t ¢ • .: ••;:tla·a..'_·�ro4',., ···�:t'Ejjl"'I'II;;,;;!11 Ii hjm!hn4wilwilii ,! $' fiiii@ ;, :;;
-1)._...
significant relationship did not exist in the case of parent-identified hyperactivity
and low income, but did exist for family dysfunction and failed a grade (Offord et
aL). For emotional disorders identified by the parent, family dysfunction, parent
treated for nerves, and chronic medical illness were significantly related (Offord
et al.), Finally for parent-identified somatization, family dysfunction and chronic
medical illness were significant (Offord et al.). For teacher/youth-identified
conduct disorder, family dysfunction, being male, and failing a grade were
significantly related (Offord et al.). For hyperactivity identified by the teacher or
youth, medical illness, low income, family dysfunction, and being male were
significantly related (Offord et al.). Furthermore, for teacher/youth-identified
emotional disorder, urban residence and family dysfunction were significantly
related whereas for teacher/youth-identified somatization, only family
dysfunction was significantly related (Offord et al.). As well, low income was
strongly related to teacher-identified conduct disorder in the 4 to 11 age group,
whereas it was not in parent- and youth-identified conduct disorder (Offord et
al.).
Although familial dysfunction continued to have a significant relationship
between individual disorders, it should be noted for both the disorder and
correlate familial dysfunction, parental identification was necessary. Because
the parent was identifying the disorder; and in this case the correlate family
dysfunction, it would not be unexpected to find concurrence or a positive
relationship to exist between the disorder and the correlate. Mitigating this
finding is the fact family dysfunction continued to be highly correlated with
individual disorders, regardless of informant.
Further to the issue of interrater agreement is another aspect of the OCHS
study conducted by Sanford, Offord, Boyle, Peace, and Racine (1992) on social
and school impairments. The purpose 'of Sanford et al. 's research was to
38 I
...iII)
determine the relationship between the four psychiatric disorders previously
outlined and social and school impairments. Social and school impairments
were defined "as a nonphysical limitation in functioning in a usual social role or
relationship" (Sanford et aI., p. 60). Impairment measures included: difficulties in
getting along with peers, teachers, and family; poor school work, student
performance, and below grade level functioning (Sanford et al.). According to
Sanford et aI., the findings indicated that: children deemed as disordered were
also more likely to be identified as impaired by all informants; parent-identified
disorder and parent-identified impairment were more highly correlated than
teacher-identified impairment and parent-identified disorder; teacher-identified
impairment and teacher-identified disorder demonstrated the highest
correlation; overall ratings of impairment among parents was significantly lower
than teacher or youth ratings; the relationship between youth-identified disorder
and impairment was the lowest; and there was little agreement among
informants as to whom was impaired.
Subsequently, Sanford et at. (1992) concluded it would be inappropriate to
base a community wide epidemiology study solely on parent reports; rather it is
believed a teacher has the most uniform view of the child, although variances in
informant rating of the prevalence of specific disorders have already been
demonstrated (Offord et aI., 1989). The limitations of the study, as it relates to
variances between informants, is the lack of data differentiating mother and
father responses and lack of self-report measures in the under 12 year old age
group due to the unreliability of the replies. As to why youths would identify
themselves more often as disordered than either their parents or teachers is
probably due to the more discriminating perception youth have of themselves
compared with to the nonspecific view parents and teachers have of the same
. youth. It was also found children identified as disordered were more
39 ..J
symptomatic when identified as impaired either socially or scholastically than
those children who were disordered but not identified as impaired (Sanford et
al.).
2.2.4 Native American Child and Adolescent Mental Health
As mentioned before, there is a lack of epidemiological data as it relates to
Aboriginal children and adolescents in Canada. To partially alleviate this
concern, a review of a study undertaken in the United States on the mental and
social health of American Indian children and adolescents was conducted. It
should be noted that there are obvious similarities between the socio-economic
demographics of the American Aboriginal population and the Canadian
Aboriginal population and that cultural boundaries are not the same as national
boarders. Furthermore, when indicators ot mental and social health are
explored, elevated prevalence rates are found to exist in the American
Aboriginal population, like their Canadian counterparts.
Although the American study focused on Aboriginal children and youth
residing in the United States, the only participants were Native American youth
in grades 7 to 12 (Blum, Harmon, Harris, Bergeisen, & Resnick, 1992). Even .
though the sample is not representative of the Native American child and youth
population, it would not be surprising to find prevalence rates which confirm the
finding that Indian adolescents are approximately at five times greater risk for
emotional disorder in comparison to the same age American population (Beiser
& Attneave, 1982; LaFromboise, 1988). Finally, the instrument used to collect
the data in the American survey is a revised version of the Adolescent Health
Survey, with 162 items addressing dimensions of health.
40
,1M' ....)
2.2.4.1 Statistical overview
Having recognized the limitations of the survey, it will be necessary to
analyze the statistical trends to see if similar indicators of mental health are
consistent across national boundaries. Like their Canadian counterparts,
American Indians have: higher infant mortality rates at 62.7 deaths per 1,000
live births compared with the U.S. average of 10.1 deaths per 1,000 live births;
mortality rates among 1 to 14 year olds which are 1.8 times the U.S. average,
with accidental deaths the leading cause of death at 27%; a higher suicide rate
of 14.5 per 100,000 compared with the U.S. average of 11.4 per 100,000; a
combined life expectancy for both sexes at 71.1 years of age which is 3.3 years
less than for the U.S. white population; a younger population with the proportion
of the population represented by 15 year olds or less at 32% compared with
23% for the national average; a declining elderly proportion of the population,
represented by those who are 64 years of age or older, at 5%, compared with
the U.S. average of 11 %; over 28.2% of their population living below the
poverty level compared with 12.4% for the total U.S. population; 13% of their
labour force unemployed, compared with the U.S. average of 6.5%; lower levels
of education with 55.4% graduating from high school compared with the U.S.
average of 66.5% (Indian Health Service, 1992).
It is clear from an examination of the socio-economic demographics and
indicators of social and mental collected by the Indian Health Centre that
American Indians find themselves in a less advantageous situation than the
American population as whole. Even though the rates of incidence and the
actual percentages vary, it is clear that similarities between the social conditions
facing American and Canadian Indians obscure national identity and lend
commonality to the social and cultural milieu of Native North Americans.
41 J-
2.2.4.2 American Indian Health Survey
Within the socio-economic and cultural milieu described, the American
Indian-Alaskan Native youth health survey was conducted in 1989 on American
Indian teenagers. The purpose of the study was to "assess risk behaviors,
health problems, worries and concerns and resiliency-promoting factors among
American Indian-Alaskan Native adolescents" (Blum et aI., 1992. p.1637). With
these objectives in mind, the survey was administered to nonurban schools
encompassing eight Indian Health centres to a total sample population of
13,454 American Aboriginal youth in grades 7 to 12. The categories the study
explored were the school environment, physical health, abuse, emotional
health, and sexual behaviours. It was found that like mainstream American
youth, the majority of Aboriginal American youth are not faced with significant
health risks, enjoy school, don't report high rates of substance abuse, are not
suicidal, and do not report high levels of abuse (Blum et al.).
What is significant about the results of the survey, as administered to this
special is while the of the and adolescents population, majority .J youth surveyed
are not at-risk, a sizable portion are at-risk, and this portion is greater than the
U.S. national average. Higher rates of emotional distress are evident among the
at-risk group where there is a growing sense of hopelessness, a general
perception of poor health among those who are deemed at-risk, and an
increase in drug and alcohol consumption among Aboriginal males in the
survey group (Blum et al., 1992). Regarding the limitations of the study, Blum et
al. state the estimations provided are conservative at best, as the sample
population was drawn from a school based setting where those who are non
attenders, and perhaps most at-risk, are not represented. As well, the sample
population did not include urban areas or rural schools where the Aboriginal
population is below fifty percent (Blum et al.).
42
------J.
In order to address these problems, Blum et al. (1992) suggest the
development of culturally appropriate health promotion efforts within the context
of community development. Thus according to Blum et al.:
Culturally appropriate not only refers to cultural considerations that emanate from Indian-white differences, but also variations in health concerns and health promotion strategies at the tribal level. Strategies that build on the strengths of community identity and culture are more likely to succeed than those imposed externally. (p. 1643)
In a similiar vein, the current state of American Indian mental and social health
conditions has led LaFromboise (1988) to conclude the causes of elevated
prevalence rates in American Aboriginal populations are stress related due to
the forced acculturation of American Indian people resulting in the
unsatisfactory adjustment to the social psychological conditions present. Thus
according to LaFromboise, research on Aboriginal American populations needs
to redress:
... [the] severe imbalance in favour of studies that focus on pathological disorders of American Indians to the neglect of investigations of milder transient problems and of research on familial or sociocultural antecedents of psychopathology. The most glaring gap, however, is the failure to examine the effective strategies currently employed by American Indians for coping with numerous stressors. (p. 388)
Given the conclusions drawn by Blum et al, (1992) and LaFromboise (1988),
it should be realized epidemiological studies in themselves are insufficient to
explain psychopathology within a school age population if they do not concern
themselves with other aspects of the child's development as it relates to their
intellectual and academic growth. In this respect, the OCHS and the American
Indian Health Survey are lacking, as the cognitive, intellectual, and academic
functioning of the child were not considered in the studies' design. Although
some very valuable information from the OCHS and the American Indian Health
Survey was garnered, as it relates to prevalence of psychiatric disorders,
43
� I - _J-
social/school impairments, social/mental health, and the correlates of the
disorders, no information was provided about the relationship between the
intellectual, cognitive, and academic functioning of children with diagnosed
psychiatric disorders or behavioural and emotional disaffection. For the
purposes of this study, it will be necessary to consult the literature to determine
what if any relationship exists between the aforementioned factors and
behavioural and emotional problems in children and adolescents. In so doing, a
better understanding of the at-risk factors which predict failure in life and in
school will be obtained.
2.3 Variables Affecting Behaviour in Children and Adolescents
Children and adolescents demonstrate many different kinds of behaviours.
The magnitude and the chronicity of these behaviours differ from child to child.
Many of the terms used to describe emotionally disturbed (EM) or behaviourally
disordered (BO) children by educators are also used to describe students who
are either learning disabled (LO) or mentally retarded (MR) (Ysseldyke &
Algozzine, 1990). It is also generally agreed that students who are
behaviourally or emotionally disordered score lower on intelligence tests and
do not do as well academically, although their abilities do span the entire range
(Hallahan & Kaufmann, 1994; Goldstein, 1986; Tramontana, Hooper, Curley, &
Nardolillo, 1990; White, Moffitt, & Silva, 1989; Worland, Weeks, Janes, & Strock,
1984; Ysseldyke & Algozzine, 1990). In terms of the present study, it is important
to understand what if any casual relationship exists between the previously
described variables. Are students manifesting problem behaviours in the school
more at risk for school or academic failure than their peers? If so, then the role
of educational programming is paramount for the delivery of services to address
the needs of the school age populations served.
44 j
2.3.1 Special Education Students and Delinquency
In a study conducted by Fink (1990), delinquency was defined as problem
behaviours which were described as encompassing disciplinary infractions in
school and serious delinquent activities. Disciplinary infractions in school
included measures of tardiness, skipping class, not completing work, being
unprepared for classes, fighting and stealing, plus delinquent activities which
included criminal acts, drug involvement, and court contacts. Although Fink's
definition of delinquency encompasses problem behaviours outside of the
classroom and school setting, valuable insight into the predictive path of
delinquency can be ascertained. Fink, in consulting the literature to determine
what models or predictive paths have been developed, determined that three
theories or models have been seen to dominate the research literature. These
models are: the school failure theory, the susceptibility theory, and the
differential treatment theory. The three models are described by Fink in
accordance with her own research on a special education subject sample.
According to the susceptibility theory, the characteristics of the
exceptionality, behavioural or emotional problem, cognitive deficit, or attention
deficit leads directly to juvenile delinquency. In the school failure model, the
characteristics of the learning disabled student lead to school failure which in
turn results in juvenile delinquency. The differential treatment theory suggests
the characteristics of the learning disabled child lead to school failure with the
labelling of the child or adolescent by the criminal justice and/or educational
systems resulting in discriminative responses and more punitive punishment
being directed towards the child or adolescent (Fink, 1990).
Although Fink's (1990) research does not demonstrate the relative effects of
the mechanisms hypothesized in the different models, the findings of Fink's
study can be interpreted in relation to the school failure model. What Fink did
45 confirm is that special education students are more at risk for delinquency, as defined by problem behaviours, than non-special education students.
Furthermore, learning disabled students reported less interpersonal competency and more negative peer influence, but did not yield higher scores on impulsivity than regular education students (Fink). Fink concluded the school failure model best describes her research because:
Analyses indicate significant differences between LD and non-disabled students on serious delinquent behavior and on in-school referrals for troublesome behavior. Both MR and LD students differed from non special education students on measures of school punishment, personal and social development, social bonding, and school performance. (p. 71)
For the purposes of this study, Fink's findings are significant. Because teachers view LD students as being more troublesome in the school, it would follow that teacher-completed ratings of student behaviour should be indicative of higher rates of in-school problem behaviours associated with the LD student.
2.3.2 Protective Effects of Intelligence on Delinquency
Further corroboration of Fink's (1990) study is provided by White et al. 's
(1989) research in which the protective effects of IQ, as measured by the
Wechsler Intelligence Scale for Children-Revised edition (WISC-R; Wechsler,
1974), were explored in relation to high risk for juvenile delinquency. In White et al. 's (1989) study, delinquent status was assessed at age 13 using the child completed Self-Report Early Delinquency (SRED; Moffit & Silva, 1988), the parent's report of socialized aggression on the Revised Problem Behavior
Checklist (RPBC; Quay & Peterson, 1983), the teacher's report on the antisocial subscale of Rutter's Child Scales (Rutter, Tizard, & Whitmore, 1970), combined with police contacts which were used to corroborate and compliment the youth
self-reports.
In this investigation, students were randomly selected and observed
46 r •
longitudinally and grouped according to the following four categories: group 1
were classified as high risk with a delinquent outcome; group 2 were classified
as high risk with a non-delinquent outcome; group 3 were classified as low risk
with a delinquent outcome; group 4 were classified as low risk with a non
delinquent outcome (White et at., 1989). Children's antisocial behaviour was
rated by both teachers and parents at 5 years of age, with high-risk students
determined through the use of local norms, meaning the top third of the
distribution for their respective gender (White et aLl. Delinquency and its
ensuing chronicity, "was defined as continuous involvement in delinquent
behavior across ages 13 and 15" (White et aI., p. 720). As a result of White et
al.'s research to test the replicability of the protective effect of la against future
criminality, it was found:
... in both low-risk and high-risk groups of girls and boys, an average or better la is associated with failure to demonstrate relatively serious and
stable delinquent behavior in adolescence.... Although some support existed for the hypothesis that delinquent girls would have significantly . lower las than non-delinquent girls, it did not appear that the delinquency main effect was as marked in girls as it was in boys. One possibility is that girls may be generally more intellectually resilient to deviance (e.g., they have fewer learning disabilities), but other factors such as family disadvantage or discord may be more important for girl's delinquency. (p. 723)
The protective effect of IQ is apparent in both low-risk and high-risk children in
that an IQ which is average or above average is not associated with serious and
stable delinquent behaviour in adulthood, with a very high la in boys, even
those at-risk, helping them avoid becoming delinquent and experiencing failure
in life or at school (White et al.), However, White et at. reach an important
conclusion regarding future research on delinquency and intelligence as it
relates to females, in that familial factors, like family disadvantage and/or
discord, may indeed prove to be better predictors of deviant behaviour than
intelligence among female children and adolescents.
47
r � I I ------��. ,------2.3.3 Familial Processes
There is left little doubt, as evidenced in current research, that the family is
the primary socialization agent for children from 3 to 5 years of age (Snyder &
Huntley, 1990). According to Patterson, Capaldi, and Bank (1991), children
learn antisocial acts through the family prior to contact with the deviant peer
group with antisocial behaviours including noncompliance, physical and verbal
aggression, lying, truancy, stealing, setting fires and running away from home.
Furthermore, it is argued the delinquent peer group's emergence during
preadolescence does not preclude antisocial behaviour in early adolescence
(Patterson et al.), It is also known that processes leading to antisocial behaviour
in early childhood can continue to operate into adolescence and late adulthood
(Snyder & Huntley).
Because of this, the family is the primary agent in coercive training. In these
dysfunctional families, it is usually poor parental discipline practices that lead to
the development of noncompliance and antisocial behaviour in the child
(Patterson et al., 1991). As the child is socialized in this type of environment, it
has been found the antisocial behaviours of the child begin to escalate and
move from noncompliance to higher amplitude behaviours like hitting and
temper tantrums (Patterson et al.). But not all children develop antisocial
behaviour and there are a number of familial factors which have been identified
as causal in the development of coercive behaviour in the child (Snyder &
Huntley, 1990):
Early development of antisocial behavior primarily occurs in family interaction. A number of converging factors interfere with family interactional processes requisite to the development of self-regulation and cognitive and social skills: difficult child temperment, unskilledlirritable parents, socio-economic disadvantage, parental stress and marital discord. Although not all antisocial children go on to become antisocial adolescents, the products of this early training (parental and peer rejection, school failure, association with antisocial peers) put children at risk for a continued drift into deviance. (p. 205)
48
______--��T- __ __J
Although familial processes are implicated in the development of antisocial
behaviour, they alone are not sufficient for its advancement. Moreover, there is
evidence the factors involved in the development and chronicity of antisocial
behaviour, in adolescence and childhood, are not necessarily the same.
Recently, Patterson et al. (1991) have advanced the early starter model which
correlates childhood coercive training practises, initiated in the family, with later
social and academic failure resulting in delinquency and possible adult criminal
behaviour. According to Patterson et aI., as the child develops, it is assumed
that coercive exchanges between the parent and child are the key ingredient in
setting the early starter process in motion. Once this begins, it seems to move
through a sequence of three stages: a) the child shows clearly identifiable
antisocial behaviours; b) the child is rejected by the normal peer group; c) the
child fails in school (Patterson et al.).
Supporting Patterson et al. 's (1991) findings is longitudinal research
conducted by Farrington (1991) on childhood aggression and later adult
violence. Part of Farrington's study relied on teacher questionnaires and it was
found that those students, 8 and 10 years of age, defined as aggressive by their
teachers, were so because they were difficult to' discipline. However, the older
students, 12 and 14 years of age, defined as aggressive by their teachers, were
so because they were disobedient, rough at play, undisciplined, quarrelsome,
and aggressive (Farrington). This led Farrington to conclude that aggression as
simply measured by teacher observation when the child is 8 to 10 years of age
is not a direct measure. According to Farrington:
" ...the causes of aggression and violence must be essentially the same as the causes of persistent and extreme antisocial, delinquent, and criminal behaviour. A great deal is known about these causes, which certainly include economic deprivation, family criminality, poor parental
child-rearing behaviour, and school failure .... [Moreover], violent and nonviolent frequent offenders tended to be drawn from low-income large sized families, in poor housing; that they tended to have convicted parents and delinquent siblings; and that they were exposed to harsh parental 49 discipline, parental disharmony, and separations from parents; and that they had low intelligence [non-verbal IQ assessed at 8-10 years of age], high impulsivity, poor concentration, and high daring. (p. 25)
As well, Farrington's findings confirm the significance of non-verbal reasoning
as a variable which correlates with future antisocial, delinquent, and criminal
behaviour. For those adolescents not identified as antisocial during elementary
grades, Patterson et al. found that they tended not to develop into chronic
antisocial adult offenders due to higher levels of social skills. Nevertheless,
these adolescents can still develop antisocial behaviour, and even though
familial processes were not coercive during childhood, parental stressors, later
in life, can affect the development of antisocial behaviour during adolescence
(Patterson et al.). Moreover, Snyder and Huntley (1990) contend that:
Changes in marital, social, vocational and emotional functioning or transient stressors that constitute the parents' own developmental context, may diminish the consistency and skill with which they monitor and discipline their children, and constructively engage them in problem solving and mutually satisfying activities. (p. 209)
Subsequently, it is recommended the earlier remediation occur for the late
starter delinquent, the greater likelihood of success (Snyder & Huntley, 1990).
Likewise, it has been suggested early starter delinquents be also targeted in
adolescence for remediation so as to reduce the possibility of becoming chronic
offenders (Patterson et aI., 1991). Clearly, familial processes are causal in the
development of antisocial behaviour in childhood and adolescence and are an
important component in assessment and intervention programs. What needs to
be determined then is how assessment can delineate those elements deemed
as causal, which in turn, can be linked to the appropriate intervention program.
50
� I ------�" ,------j
2.3.4 Familial Factors. School Behaviour. and Learning Disabled Students
Toro, Weissberg, Guare, and Liebenstein (1990) conducted a research
project which was a comparative study of LD children and children without
learning disabilities (NLD) along the variables of social problem solving skill,
school behaviour, and family background. Social problem solving was
determined through the Open Middle Interview (OMI; Polifka, Weissberg,
Geston, Flores de Apodaca, & Picolli, 1988) scored along the factors of
alternative solutions, solution variants, and irrelevant responses. School
behaviour was determined through the Child Behavior Rating Scale (CRBS;
Weissberg et aI., 1981) which yielded factor scores on acting out problems, shy
anxious problems, learning problems, frustration tolerance, adaptive
assertiveness, peer sociability, and both overall and global adjustment (Toro et
aI., 1990). Limitations of the study as recognized by Toro et al. relate to the
selection of LD students, as defined by school placement, and the use of
teacher ratings for the determination of school behaviour and family
background. Toro et al, suggested teachers, who were not blind to student's
status (LD or NLD), might have provided responses based on assumptions
regarding the LD's familial background.
However, what the research results do indicate is that deficits in the
cognitive and academic domain as well as social and behavioural areas was in
part due to inadequacies in the home environment (Toro et aI., 1990).
Specifically, economic difficulties, lack of educational stimulation, and familial
problems were all significantly related to LD students, with LD students not
being necessarily over disruptive but rather deficient in less visible social skills
(Toro et at.). It is clear from Toro et ai's findings greater attention needs to be
placed on the role of the family. Where previous research seems to indicate that
social and behavioural differences are the result of cognitive deficits, "it is also
51
r • possible that the cognitive deficits could be the result of the social deficits, or of
poor early environmental conditions" (Toro et aI., p. 118).
2.3.5 Socio-economic Status, Behaviour, and Intelligence
Further to the issue of environmental effects and their relation to behaviour
is a British study conducted by Kniveton (1987) on imitative misbehaviour in the classroom. In this study, imitative behaviour was determined through
behavioural observations of students who had previously viewed a video tape of a peer model displaying mischievous and naughty behaviour. According to
Kniveton:
The peer model was shown to throw and kick and stamp on a pile of paper cups; to hit some balloons and then stamp on them bursting a number of them; to punch a large toy panda and throw ltat the wall; to shoot a gun with 'bullets' at a picture of a man on the wall; and to knead Playplex viciously. Basically the 'misbehaviour' shown was of the type which could be expected of a boy being naughty during unsupervised play periods. (p. 164)
In Kniveton's study, it was found that students of low socio-economic status
(SES) were more likely to imitate aggressive behaviour after viewing the videotape of the peer model than those students of middle income. The
students were also tested for intelligence using the Wechsler I ntelligence Scale
for Children (WISC; Wechsler, 1965a� and no significant relationship was found to exist between imitative misbehaviour and intelligence (Kniveton). This study
then does not discount the relationship between intelligence and classroom
behaviour, but rather states that intelligence is not significantly related to
imitative misbehaviour, while low SES is strongly related. A possible
explanation as to why SES is significantly related to imitative misbehaviour is
the subordinate status afforded the low SES child in the classroom by peers of
higher SES; hence the low SES child acts out to gain attention and win favour
among their classmates.
52
r- r Low SES has also been confirmed as significantly impacting the development and subsequent intelligence and behaviour of students (Dumaret,
1985). A study conducted by Dumaret involved siblings raised in contrasting environments due to adoption and examined both the long and medium term effects of social environment on intelligence and behaviour. The instruments used to collect the data on intelligence included the Collective Scale of
Intellectual Level (ECNI; INED, 1973) and the WISC. Behaviour was determined through the transcribing of all school records, the behaviour "B" scale developed by Rutter (1967), and a questionnaire designed by the research team to measure school adjustment and behaviours. The questionnaire included factor scores for school failure, school failure and serious problems, school failure at the third primary level, gifted, and well adjusted (Dumaret). The results of the study led Dumaret to conclude:
[an] increase in IQ and diminution of scholastic failures for the adopted children, diminution of IQ and increase in scholastic failure for the others show that environment has important effects. Adoption played a dynamic role, permitting these children to develop their intellectual resources
thanks to a favourable social, cultural and familial environment. ... [However], the accumulation for failures is the greatest for children raised in a poor environment under conditions which have led to their being abandoned to institutions or foster homes. (p. 573)
Dumaret's findings underscores the role in which environment can affect children and youth placing them at-risk. To further understand th� relationship between SES, intelligence, academic achievement, school behaviour, and psychopathological risk, Worland et at. 's (1984) path analysis will be discussed as a model was delineated to explain the degree of association between the aforementioned variables.
In Worland et al. 's (1984) inquiry, the measures used to determine intelligence were either the, WISC-R or the Wechsler Adult Intelligence Scale
(WAIS; Wechsler, 1955). Classroom behaviour was assessed by the Pupil
53
"II' ",i ".:1 ...... )
Rating Form (PRF; Watt, Grubb, & Erlenmeyer-Kimling, 1982) composed of
twenty eight items which provide factor scores for scholastic motivation,
extroversion, harmony, and emotional stability. Academic achievement was assessed using the Wide Range Achievement Test (WRAT; Jastak & Jastak,
1978). The sample was chosen from a longitudinal study being conducted by
the authors and involved children whose mean grade was between Grades 7
and 8 and defined as low, moderate, or high risk status for psychopathology. In
path analysis, a causal model cannot be proved, rather a model is specified in
mathematical terms and then the data are used to either verify or dispute the
model's design. In so doing Worland et al. developed:
... a model in which the casual relationships were from intelligence,
through academic achievement, to classroom behavior... The direction of the relationship suggests that the behavior these children exhibited in school was the result of academic achievement rather than a cause." (p.449)
When SES and risk of psychopathology, as determined by parental
psychopathology, were factored in, it was found "both SES and Risk were
correlated independent variables that simultaneously and similarly affected
intelligence and (through intelligence) the other variables [academic
achievement and school behaviour]" (Worland et aI., p. 448). It should be noted,
that for the purposes of our dlscussion, classroom behaviour in the Grade 7 to 8
sample described is an effect of academic achievement rather than the cause.
2.3.6 Summary
What demographics, mental health indicators, and epidemiological studies
do is provide a descriptive picture of the society and/or population under study.
Research uncovering the relationships between la, behaviour, academic
performance, and SES lend support to predictive models which articulate what
places children and adolescents at-risk. However, both these type of studies are
54
- _""",.,.-, r I somewhat limited in their value, unless combined with the theoretical
underpinnings of psychopathology across cultures which can assist the
caregiver, clinician, and educator in delineating the cultural factors which
interact not in isolation, but in concert with the social, economic, behavioural,
intellectual, and cognitive indicators outlined. The review of literature then, will
examine issues pertaining to psychopathology across cultures in an attempt to
explore the variances in culture which might complicate the assessment
process.
2.4 Psychopathology Across Cultures
It is known that cultures vary worldwide and every culture is composed of
beliefs, values, attitudes, and behaviours, to a greater or lesser degree, intrinsic
to itself (Ramirez & Castaneda, 1974). When discussing culture, according to
Berry, Poortinga, Segall, and Dasen (1992), it is important to realize that culture
is not civilization, as all groups of people have culture and therefore, the
civilized, primitive debate is one which is to be avoided; secondly, culture is not
'high culture' referring only to the arts, but rather it is all products of human
evolution; thirdly, culture is not to be equated with society, as society refers to
people whereas culture refers to a way of life; and finally, there is a distinction to
be made between explicit culture, the observable behaviours and products
found in a group, and implicit culture, which urefers to the organizing principles
that are inferred to lie behind these regularities on the basis of consistent
patterns of explicit culture" (p. 168).
When examining cognition and cognitive processes and how they intertwine
with personality, it should be noted the behaviours manifested by an individual
are part of a triadic relationship between an individual's cognition in
development, their behaviours, and the environment (Kaufmann & Kneedler,
55
.I� '·�;�·f" _.J ------�� ;------... 1981). Seen in this context, the influence of culture on the normative behaviour
and cognitions of a group can be distinct to the society in which they occur,
depending on the forces operating within and between cultures in contact.
As if the issue of assessment was not complex enough, it should be realized
that fundamental to the discussion surrounding culture and the assessment of
disorders is one of definition. Variations in the expression of a disorder or
behaviour can be culturally induced, specific to the culture in question, or a
result of acculturation leading to psychological change.
2.4.1 Culture Bound Syndromes
According to Simons and Hughes (1993), Greenlandic Inuit have a
syndrome specific to their culture labelled as kayak angst, which is described as
an intense fear of capsizing and drowning on the ocean while in a kayak. Berry,
Poortinga et al. (1992) use the term culture bound syndromes to describe
disorders which are culturally induced, like kayak angst:
The hallmark of all these syndromes is their exotic (to Western observers) qualities; each is usually given an interpretation within terms of their own culture. The issue for cultural psychiatry is whether they are also comprehensible within a universal framework of psychopathology. (p. 364)
The question researchers are now exploring is the extent to which culture
bound syndromes are unique expressions of a syndrome specific to a culture,
and the degree to which the expression of a culture bound syndrome is not a
local expression of a particular universal psychosis but rather an indigenous
expression of a more commonly understood psychological principle. Of course
the limiting aspect of such research is the degree to which the evaluation occurs
within the framework of Western psychiatric nosology, jeporadizing the
ecological validity of the study.
56
.------£� 1------)
2.4.2 Psychological Change and Acculturation
Psychological change in the behaviour of the individual and the group
cannot be understood unless the effects of acculturation are also realized
(Berry, Poortinga et aI., 1992). Within the context of First Nations, it is important
to remember that the intensity of the cultural confrontation was usually dictated
by outside forces (Adams, 1989; Ecceles, 1988; Kirmayer, 1994b). There are
essentially six factors involved in how acculturation might affect the group: (a)
environmental alterations, including urbanization, relocation, and pollution; (b)
biological changes in diet and disease; (c) political structure changes, in terms
of dominant and subordinate power structures; (d) changes in the economic
system, including subsistence and labour market economies; (e) cultural
changes referring to religion, language, way of life, and institutions; and (f)
social structure changes, meaning between and within group changes (Berry,
Poortinga et at.: Kirmayer).
Factors identified as influencing psychological change as a function of
acculturation include; perception and cognition, personality, identity, attitudes,
and acculturative stress (Berry, 1980). Thus psychological acculturation is the
result of individual changes due to contact with another culture and occurs
when one participates in the acculturation process (Berry, Poortinga et aI.,
1992). The distinction between group acculturation and individual
psychological change is an important one for two reasons. Changes which
occur at the group level in a population can result in economic, social, or
political change and effect the individual in terms of their values, attitudes, and
beliefs (Berry, Poortinga et al.). Secondly, because not everyone reacts in the
same way to the processes acting on the larger group or population, individual
psychologies need to be considered.
According to Berry (1980), the degree to which the individual will experience
57
, psychological change is a function of the degree of acculturation which has
taken place. For our purposes, issues relating to personality, attitudes, and
acculturative stress are relevant. In terms of personality, identity, and attitudes, a
crisis or conflict arises within the individual which causes one to reach a critical
point during the medium or transitional phase in acculturation, causing a shift in
personality, identity, and attitude. Consequently, the individual will either return
to their traditional identity, rejecting the attitudes of the dominant culture and
maintaining a personality which is quasi-traditonal in composition (Berry; Berry,
Poortinga et al., 1992). Similarly, an individual may become assimilated,
rejecting the attitudes of their traditional culture, developing a new cultural
identity and thereby, taking on a personality which is similiar to the personality
traits reinforced in the new culture (Berry; Berry, Poortinga et aL).
2.4.3 Acculturative Stress
Behaviour change as a result of acculturation can either result in
behavioural shifts or acculturative stress (Berry, Poortinga et al., 1992). When
behavioural shifts occur, there is qualitative continuity in the change. Within the
parameters of behavioural shifts are changes in personality, identity, and
attitudes. Acculturative stress is more debilitative and is similiar in notion to
culture shock and cultural disintegration (Berry, 1980; McShane & Berry, 1988;
Berry, Poortinga et al.). Acculturative stress has also become the term used to
describe behaviours which are pathological and disruptive in nature. Thus
according to Berry, Poortinga et al.:
... [acculturative stress] refers to new phenomena that often accompany acculturation, which appear to result from psychological conflict and social disintegration such as an increase in homicide, spouse abuse, or a decline in mental health status; this type of consequence has been termed acculturative stress and is characterized by a qualitative change in the life of an individual or community. Once again it is possible to challenge the distinction; after all, homicide, aggression and neglect are present in most societies prior to acculturation, but we frequently 58
J,.-..... "'''',; '1 ...... " ...... , I ------��-.------1------
...)
encounter new forms as well as new rates, such that rather than being regarded as deviant and negatively sanctioned in the group they become common. (p. 280)
Factors identified as affecting acculturative stress are the mode of
acculturation, phase of acculturation, nature of the larger society, characteristics
of the larger group, and characteristics of the acculturating individual (Berry,
Kim, Minde, & Mok, 1987). Because the degree of acculturating experience can
vary among individuals, with the number of stressors varying from a few to
many, the degree of acculturative stress experienced can either be high or low
(Berry, Poortinga et al., 1992). It is important to realize then that not always do
negative effects in terms of disorganization and disintegration occur during
acculturation; rather this process is determined by the quality and features of the
relationship between the cultures in question (Barry, Wintrob, Sindell, &
Mawhinney, 1982).
Nevertheless, nomadic people, who were hunters and gatherers, such as
Aboriginal people in pre contact North America, usually suffered more negative
consequences as a result of acculturation than those who were sedentary prior
to contact (Berry, Poortinga et al., 1992). Changes in status, relative to the new
entry status one receives when being acculturated, and one's age and gender
have also been documented as factors affecting the degree of acculturative
stress one will experience (Berry, Poortinga et al.), As well, since psychological
variables may effect the mental health of individuals experiencing acculturation,
it is necessary to determine prior intercultural knowledge, whether previous
contact was voluntary or not and the initial attitudes of those about to become
involved in the process towards acculturation (Berry, Poortinga et aL).
59
,:'''" r .�' � I I l I
,....)
2.4.4 Somatization and Social Response to Distress
Among indigenous cultures, there is a tendency for disorders to be
expressed in somatic presentations (Kirmayer, 1989). In Westermeyer's (1993)
opinion, this is probably so because some cultures endorse somatic
presentations and a patient may be reluctant to trust a clinician with personal
information other than somatic symptomatology, especially if the assessment is
not occurring in the culture of the patient.
But somatization is not just common to indigenous cultures. The DSM IV
(American Psychiatric Association, 1994) has a specific classification for a
syndrome described as somatization, that is designated as exhibiting symptoms
concerning bodily affect. There is also evidence to suggest that the presentation
of somatic symptoms occurs more frequently in indigenous cultures when
anxiety and depression co-occur (Kirmayer, 1989).
During social change, acculturative stress occurs and has been used "as a
general term to refer to those individual states and behaviours that are mildly
pathological and disruptive, including those problems of mental health and
psychosomatic symptoms" (Berry, 1980, p. 261). Kirmayer (1989) further
explains that more traditional cultures have a wholistic world view where the
dualistic separation of mind and body, common to Western civilization, has not
occurred. Western civilization's concept of the self is one of individuality and
autonomy, in which independence and separateness are sought (Berry,
Poortinga et aI., 1992). As a result, traditional cultural values have been
supplanted in favour of a dualistic approach to one's concept of self, which
discriminates between disorders in terms of bodily and mental affect. Rather
than attempt to explain distress as an interpersonal or intrapsychic conflict,
traditional cultures will refer to the same distress in terms of bodily functions
(Kirmayer).
60 j
2.4.5 Deviance
Labelling theory argues, in both its weak and strong forms, that the
presentation of distress in an indigenous culture can either be exacerbated or
mitigated through the labelling of deviance. Subsequently, it is recognized that
there are three processes which can be studied independently but should be
regarded as representative of a whole when studying deviance. The
determinants, as to whether or not an expression of distress will be labelled
deviant, are dependent on the cultural norms of the society, the size and
integration apparent within the society, and the degree to which stigmitization
will or will not occur (Kirmayer, 1989).
2.4.5.1 Cultura"y labe"ed deviance
Having alluded to the concept of culturally inappropriate behaviour being
labelled deviant, it is now time to examine deviancy and its relationship to
culture. Simply stated, distress, in order to be labelled deviant and therefore,
ascribed a descriptive category within a culture, has to violate the normative
standards of the culture's world view (Kirmayer, 1989). As we", clinical
diagnosis and intervention surrounding a particular distress is reliant on the
culture's implicit recognition that a disorder exists (Kirmayer; Robin & Spires,
1983). A disorder would not be able to present itself if there were not a culturally
defined label for it and, therefore, if it is not labelled or recognized from within a
culture, any attempt to assess or remedy the disorder might meet with frustration
and failure due to the lack of cunurally relevant definitional criteria.
61 2.4.5.2 Societal integration and deviance
The second issue raised as having a significant impact on the labelling of distress has to do with the degree of integration characteristic in any culture.
Kirmayer (1989) claims that social cohesiveness in small scale societies translates itself into a fairly egalitarian world view where nominal deviance is accepted, whereas gross deviant behaviour is mitigated due to the high levels of cooperation in such societies resulting in strong pressures on the individual to conform. Brody (1987) in his discussion of northern arctic hunting culture's mechanisms of authority addresses the issue of social control and the presentation of deviant behaviour rather succinctly:
Part of the answer to this question [of the presentation of deviance] lies with the hunter's mobility. Individuals or groups that cause trouble are avoided. This means that they will find themselves alone in a world where survival depends on a network of interdependence. The threat hanging over the head of all trouble-makers, therefore, is that their friends and relations will move away from them. In fact, northern hunting peoples are tolerant and generous to their fellows. A man who refuses to work, for example, will be fed: the sharing of meat is unconditional. But persons whose refusal to work begins to constitute a liability, and so jeopardizes others' success, are left to their own devices. In this way, the hunter's preparedness to shift camp establishes a strong deterrent as well as an actual, if indirect reprisal. By this mechanism deviance is rendered uncommon and, at the same time, can be easily dealt with. (p. 125)
However, there has been rapid acculturation in the arctic, and as a result, the traditional mechanisms of social control are no longer as viable. In their place, the Canadian criminal justice system has become the dominant means of social control by which deviant behaviour is defined and dealt with. This does not mean the higher rates of violence, incarceration, homicide, suicide, and substance abuse (Report of the Steering Committee Medical Services Branch,
1991) are solely attributable to the lack of traditional SOCial control, but rather that it is one of many factors which contribute to the current malaise facing
Aboriginal communities in the Canada.
62 ------�
2.4.5.3 Stigmitization
The third issue regarding the presentation of deviant behaviour in the cross
cultural context has to do with stigmitization. When behaviour is labelled as
deviant, negative effects in the form of stigmitization become present. Further to
this discussion, stigmitization can manifest itself in various ways including job
loss, social isolation, and loss of social status, resulting in an exacerbation of
the emotional disorder in question' and reduced prognosis for cure (Kirmayer,
1989). Thompson, Walker, and Silk-Walker (1993) raise the issue of
stigmitization as it relates to American Indians and Alaska Natives, and they are
careful to caution about making generalizations which are applicable to all
Aboriginal populations regarding its prevalence. According to Thompson et aI.,
the extent to which stigmitization presents itself in Aboriginal cultures is
dependent on the following casual factors: the distinction made in the culture
between mental and physical disorders (somatization); the degree to which
acculturation and the subsequent learning of the dominant belief system has
taken place; and the degree to which the society has urbanized. Regardless of
what casual factors are attributable to the stigmitization of deviant behaviour, it
is Thompson et al. 's thesis that the clinician needs to have knowledge of the
patient's culture, the patient'S attitude towards mental illness, and the patient'S
attitude towards the delivery of mental health services in his/her community in
order for assessment to be culturally fair. Similarly, it is Kirmayer's contention
that the culturally induced concept of the self is the mediating factor involved in
stigm itization:
Where the person is conceived of in terms of the family or larger social unit, rather than an isolated individual, the stigma of illness affects the entire unit and demands a collective response to reintegrate or expel the deviant member. Where the person is defined in terms of an autonomous rationale agency, as in the West, stigma may be less a. shared social problem and various forms of marginality, short of complete extrusion, can evolve. It is in this context that biomedicine has developed
diagnosis for deviant behaviour .... (p. 334) 63
.. _Ir-"" J ------��- .'------Therefore, in order to define a disorder, it is necessary to view the patient in terms of how the culture defines the concept of self and the subsequent degree to which the disorder is suppressed, punished, or made not plausible by the culture in question (Kirmayer).
2.4.6 Schizophrenia
It is generally agreed that there is some casual evidence implicating a relationship between schizophrenia and heredity. It has also been recognized though, that schizophrenia can be induced by cultural experiences. Likewise, the risk of developing schizophrenia is exacerbated through cultural factors such as the processing of information, complexity of messages received, and delineation of ambiguity (Kirmayer, 1989). Culture is implicated in the definition, recognition, acceptance, and symptomatology of the disorder, including its course and prognosis (Kirmayer). Kirmayer purposes the factors effecting the course and prognosis of schizophrenia vary across cultures and where the disorder is not stigmatized, support mechanisms within the society and family tend to mitigate the disorder and allow a greater prognosis for cure, while
"cultures that ascribe moral blame to the psychotic person with no opportunity for redress, may promote chronic disability" (p. 334).
However, caution should be exercised in ascribing schizophrenia as a universal disorder with prevalence rates and modes of expression culturally induced. This is because any and all cross-cultural research into the psychopathology of schizophrenia has been bound by Western psychiatric nosology and sampling techniques that have not been representative (Berry,
Poortinga et aI., 1992). Furthermore, according to the Report of the Steering
Committee Medical Services Branch (1991):
It is generally suspected that a significant number of Aboriginal people are misdiagnosed for reasons such as a proneness to interpret cultural behaviour as schizophrenic symptomatology, or to miss a dual diagnosis. 64
_,-- -::;:
.. "�:1 1 ------This, together with long-term holding of Indian patients in psychiatric hospitals because no community care is available for them, contributes to a disproportionately high number of Indian patients in lengthy psychiatric hospitalization. (p. 12)
2.4.7 Depression
Depression has been the subject of a number of studies in cross-cultural psychopathology. It is known, that like schizophrenia, both family and societal response can either mitigate or exacerbate the disorder (Berry, Poortinga et al.,
1992; Kirmayer, 1989; Marsella, 1980). As a result, when the epidemiology of depression is considered cross-culturally, prevalence rates have been found to vary. The research strategies used to uncover both treated and untreated prevalence rates, corresponding to the manifestation of depression within a culture include; culture bound syndromes, clinical diagnosis, surveys, factor analysis, matched samples, and matched diagnosis (Marsella). Research in objective versus subjective cultures has found that in objective cultures, which are characterized by an abstract language and individualization, depression is primarily affective and cognitive in form. On the other hand, cultures which are considered subjective, where language is metaphorical and living is communal, the manifestation of depression is primarily somatic in nature (Berry, Poortinga et al.). At the same time, Marsella concludes:
The more Westernized a culture, the more one can expect a picture that includes both somatic and psychological components. In contrast, depression in less Westernized cultures often does not involve psychological components. Among the major differences in depression between Westernized and many non-Westernized cultures are the absence of guilt and self-deprecation. Suicidal attempts and suicide ideation are also rare in many non-Western cultures. Given these differences, it is questionable whether Western depression can be considered a universal disorder. (p. 276)
65 2.5 Cross-Cultural Assessment
There is a growing movement among psychologists, caregivers, and educators in Canada today to develop culturally sensitive psychological instruments for the testing of minorities. This movement is partially fuelled by controversies in the United States surrounding ability and aptitude measurement. Some strategies have been suggested for improving cross cultural measurement in the behavioural and personality domains, as well as in the intelligence and aptitude domains. Research questions pertaining to the measurement of abilities and behaviour in disparate groups and whether a test measures the same "thing" in both groups have been guiding research (Irvine,
1985).
However, the field of assessment has been wrought with controversy. There are those who would advocate less use of formal assessment procedures and in the extreme, eliminate testing or quantitative measurement. If this were the case, subjective human judgment would prevail. From the other perspective, inappropriately norm referenced standardized testing is the guiding dogma.
Between these two positions lies a middle ground supported by scientific research. This controversy has partially evolved because "scientific truths and the power to change people's lives are uneasy bedfellows, and in the use of test results cohabitation becomes inevitable" (Irvine, 1985, p. 166). What Irvine advocates then is the need for greater cross cultural research in the field of testing so culturally fair assessment practices can be used with minority or culturally diverse groups.
The review of the literature then, will examine the field of cross cultural measurement. Research as it relates to cultural, linguistic, and ethnic minorities will be reviewed with specific reference to Aboriginal North Americans. Within this rubric, formal assessment practises will be discussed including the
66
l _ 411----'.·--- .;
ability/aptitude discrepancy debate and behavioural/personality measurement
in Aboriginal populations. Concurrently, the efficacy of formal assessment
procedures within cultural, linguistic, and ethnic minorities will be critiqued.
From this body of knowledge, recommendations for best assessment practices
with Aboriginal North Americans will be drawn incorporating both formal and
informal assessment methods.
2.5.1 Formal Assessment
Research in cross cultural testing has evolved along three distinct
paradigms. One direction research in testing has developed is construct
referenced meaning. In construct referenced meaning, the degree to which
cultural constructs pervade and are unique to a specific culture and the extent to
which the environment has shaped individual personality and has had an effect
on group psychology are studied.
The second area of research development in cross cultural testing relates to
criterion referenced meaning. In criterion referenced meaning, the researcher or
experimenter is trying to determine the extent to which a test made in one
culture can be used with validity in another culture. This is especially relevant
for intelligence testing as it questions the knowledge which is required of an
individual to score well on a culturally biased test. The operationalization of a
test and its use is the research question posited, and what one should be able
to do is determine the cultural distance between groups and the degree of
acculturation that has affected the individual or larger group being tested.
The third area of investigation that has influenced research in cross cultural
testing has seen the development of tests within a specific culture for use only in
that culture. This implies different tests are developed for different cultures and
are specific to the culture in which they are developed. To this end, there have
67 )-
been experimentations in behavioural testing in which translations of an
English language assessment instrument have been developed for use in a
culturally different linguistic group (Frontera-Benvenutti, 1991).
Research in personality assessment has been fraught with more difficulties
and contradictions than ability measurement. Irvine and Carroll (1980) make
this point in their discussion on testing and assessment across cultures:
There is as little, or as much, scientific support for the positive view that personalities and cultures vary together as there is for asserting that there are no co-varying differences in cognitive style or structure across cultures, or educational levels. In fact, to hold the first position is even more difficult, scientifically, than to hold the second, because theories, and consequently methods of personality research, are as diverse as the whole of psychology itself. Hence, although the quest for personality assessment across cultures may not be fraught with a" the social opprobrium that greets assessment of abilities across cultures, it is even more complicated in that many different theoretical strands intertwine in the material presented in self-report tests and in other kinds of personality assessment. The risk of social censure is smaller, but the risk of scientific opprobrium is greater. (p. 226)
2.5.2 Emics and Etics
As has been alluded to, there are essentially two major research traditions
guiding cross-cultural psychology. The first intends to explore indigenous
psychological phenomena and the extent they are related to the culture in
question. The emphasis is on the singular culture and the cultural context of
psychological processes. The second major research tradition' guiding cross
cultural psychology, and perhaps the more common within the field of
psychology itself, is the attempt to understand patterns of behaviour and
relationships across cultures so as to delineate universal patterns of behaviour
or psychological processes. Pike (1967) was one of the first to articulate this
distinction and he did so by conceptualizing the two research paradigms as
follows:
68
_ .. ,
, � ... ------
...... )
It proves convenient-though partially arbitray-to describe behavior from two different standpoints, which lead to results which shade into one another. The etic viewpoint studies behavior as from outside of a particular system, and as an initial approach to an alien system. The emic viewpoint results from studying behavior as from inside the system. (I coined the words etic and emic from the words phonetic and phonemic, following the conventional linguistic usage of these latter terms. The short terms are used in an analogous manner, but for more general purposes.) (p.37)
In making the emic-etic distinction, Pike (1967) outlined a series of principles
to explain their difference in relation to language and culture. The principles
outlined by Pike, and paraphrased by Berry (1989), are: (a) the etic researcher
examines all cultures and languages simultaneously, as opposed to the emic
research whose singular focus is on a specific culture or linguistic group; (b) etic
variables under study are discovered or known in advance, whereas emic
classifications need to be discovered; (c) etic systems can be based and
organized on a universal schema created by the researcher while emic systems
are defined and/or discovered outside of the particular system or universe of the
researcher; (d) etic descriptions are criteria formed external to the system while
emic descriptions are based on views garnered from participants who can
function in the system under study; (e) etic criteria can be absolute and are
external in relation to the system under study while emic criteria are relative to
the internal functioning of the system; (f) etic analyses are parts specific while
emic analyses are more holistic; (g) in etic studies, significant differences are
based on instrument measurements while in emic studies, variables are
different when they elicit different responses; (h) and in the final presentation of
data, etic analyses, which are the starting point, through redefinition, are
gradually described as principles which are emic in nature.
69 - __------
2.5.2.1 The combined emic-etic approach
It is not argued that one approach is more favourable than the other when
conducting cross-cultural research. Rather it should be remembered that both
have benefits and limitations, and the whole emic-etic distinction is best viewed
along a continuum. Rather than dichotomizing the distinction, and be 'blind in
one eye', the emic/etic approach allows the researcher use of both eyes. Even
so, psychology as a tradition attempts to explain psychological phenomena
using generalization and a comparative approach. In contrast, the anthropology
views the comparison of phenomena from different settings as a
false enterprise. To resolve this debate, in order for cross-cultural comparisons
to be valid, functional equivalence must be demonstrated. Berry (1989)
summarizes the inherent discrepancy by concluding:
When aspects of behaviour occurring in differing behavior settings are functionally equivalent, then a comparative descriptive framework, valid for both behaviour settings can be generated from an internal description of behavior within each setting. Only when both these conditions are
. met, may one attempt to use concepts and develop instruments to gauge behaviour within each setting. Any such attempt leads to further requirements that have to be satisfied: the instruments have to be conceptually equivalent in individuals in the two settings and the data obtained have to be metrically equivalent. (p. 726)
Although Berry uses behaviour to explain his point, it should be remembered
that similiar conclusions can be drawn surrounding the psychological constructs
and processes which are tapped in the measurement of ability and
achievement cross-culturally. In Figure 2.13, the operationalization of emics and
etics are concpetualized. Unfortunately, precise operationalization of emic and
etic terms is unlikely to occur as it would in the physical sciences, nevertheless,
the guides to cross-cultural research, as they relate to equivalence and
operationalizing emics and etics, provide a basis by which studies can be
subject to both scrutiny and scientific discussion (Berry).
70
_jW*"'''�' ,.----.,. RESEARCH CULTURE A CULTURE B ACTIVITY (OWN) (OTHER) 1. Begin research EMIC in own culture A
2. Transport to �IMPOSED other culture ETIC I 3. Discover other EMIC culture B • 4. Compare two EMIC .... � EMIC cultures A B
5-1. Comparison not EMIC EMIC possible A B
5-2. Comparison EMIC EMIC " '.' possible A 0';.· B DERIVED ETIC
Figure 2.13. Schematic Representation of Operationalization of Emics and Etics. Source: Berry, 1989; p.730.
2.5.3 Psychometric Equivalence and Cross-Cultural Measurement Strategies
2.5.3.1 Direct comparison and crude translation
According to Hui and Triandis (1985), direct comparison and crude
translation are probably the simplest and most popular methods for
demonstrating equivalence in cross-cultural research. Simple comparisons
between culturally different populations/samples administered the same
instrument can be accomplished using t-tests or MANOVA. The underlying
assumption is that the construct being measured exists and is operationalized
the same way in both cultures. Furthermore, it is assumed that scalar
equivalence is being met with the instrumentation, leading critics to question the
legitimacy of assuming that all the assumptions made are being met.
71
� I .. ------
2.5.3.2 Regression methods
Poortinga (1975) has used regression analysis to demonstrate scalar
equivalence. Essentially, regression analysis undertakes to determine if the
parameters of the criteria or construct being measured and scored are the same
for the populations being studied. Thus if scalar equivalence is said to exist
between two test scores, then the construct being measured must be related to
an external criteria (Hui & Triandis, 1985).
2.5.3.3 Co-scoring methods
The underlying assumption made when attempting to demonstrate or
improve equivalence using the co-scoring method is the presupposition that
items, and subsequently factors, have conceptual or functional equivalence. It is
argued that unless there is a strong theoretical basis for assuming the
equivalence of a construct cross-culturally, then co-scoring should not be used
as factors will always be evident, unless the variance in the data is negligible
(Hui & Triandis, 1985).
2.5.3.4 Response pattern method
The response pattern method is advocated by Irvine and Carroll (1980) to
determine if item equivalence in aptitude and ability testing exists. 8yordering
the correlated rankings of item difficulty, it is maintained that items which are
more difficult for subjects in culture A, as opposed to subjects in culture 8, are
made evident. Although the response pattern method is suitable for
demonstrating item equivalence, it does not prove or disprove that scalar
equivalence exists. However, to extend the analysis beyond the realm of
aptitude and achievement testing into behavioural or psychosocial
measurement can be problematic (Hui & Triandis, 1985). According to
Hui and Triandis:
72
• ".... i�
.... ,_.-J!i ---- ... the mean degree or frequency of endorsement of items (in personality or social psychological measurement) can be ordered for the same analysis. Unfortunately, there is no commonly held objective standard by which we can determ ine whether a correlation coefficient is sufficiently high so that item equivalence can be safely assumed. Moreover, the coefficient is only an indication of overall similarity in the relative endorsement frequency of each item. Therefore, it can be misleading if all items are higher or lower in the frequency of one culture than for another. (p.140)
2.5.3.5 Translation technigues
If there is a lack of item equivalence, then it is suggested that it may be corrected through the use of translation techniques. Presently, the translation methods offered are the back, committee, and decentering approaches (Hui &
Triandis, 1985). The most common and frequently used translation method is the back technique, where the researcher translates the items, questions, or passages into his/her mother tongue. In conducting back translation, two bilingual translators are employed. One translates the items from the language of the instrument into the mother tongue, the other, blindly, translates the items from the mother tongue into the language in which the test was originally written. The researcher's responsibility is now to determine if the two versions of the test in the original language are identical, and if not, then corrections need be made for the items in question. Although translation will ensure linguistically, that items, questions, or passages are equivalent, it would be wrong to assume
I that this method proves that the constructs are equivalent.
2.5.3.6 Internal structure congruence
As the name implies, internal structure congruence analysis, as applied to cross-cultural research and equivalence, is based on the premise that if a construct is similar across cultures and therefore, approaches universality, it should have the same or similar internal structures. Essentially there are four
73
-.-�------.. -----
distinct methods which can be used to determine internal structure congruence,
and they are factor analysis, multidimensional scaling, maximum likelihood
factor analysis, and comparison of correlation matrices. Of these four methods,
factor analysis appears to be the most popular (Hui & Triandis, 1985).
Multidimensional scaling can be applied when there is a sample from each
culture who judge the appropriateness or similarity in meaning of an
instrument's item pairs (Hui & Triandis). Statistically, multidimensional scaling
can be used to explain what if any similarity, or variance, exists in the judges
responses. Within the Canadian context, Berry and Bennett (1992) completed a
study where Cree conceptualizations of intelligence and well being were
described using multidimensional scaling analysis.
To prove maximum likelihood factor analysis, confirmatory factor analysis is
being used more frequently. Using LlSREL, a chi-square test of goodness of fit
is applied to determine if the data garnered from different groups is sufficiently
similar to be described by a unique factor model (Hui & Triandis, 1985). This
goodness of fit procedure has been termed the maximum likelihood estimation
and it differs from traditional factor analysis in that additional information is
provided through the construction of variance-covariance matrices.
If the researcher does not want to test if the factor structures are similar in
both cultures, then intercorrelation matrices can be constructed. Essentially an
"asymptomatic chi-square test can be used to test the null hypothesis that the
two sample covariance matrices are drawn from the same population matrix"
(Hui & Triandis, 1985; p. 143). Again, even though internal equivalence might
be demonstrated, it would be wrong to assume that scalar equivalence exists.
74
.• 2.5.3.7 Nomological network and validation Cronbach and Meehl's (1955) definition of construct validation was the impetus behind the concept of validation by nomological network. Nomological network validation determines if the construct is embodied in the nomothetic network in one culture as it is in another. Thus if constructs are valid cross- culturally, it follows that they should enter into the same empirical relationship (Hui & Triandis, 1985). Problematic with this procedure is the assumption that equivalence is met even though item and scalar equivalence have not been proven. Furthermore, in order to meet criteria for construct validation, it is necessary that the criteria or structures of other networks are similar. Thus by itself, nomological network validation is incomplete; rather a researcher needs to incorporate a number of strategies to improve the cross-cultural validity of measurements. 2.5.3.8 Multi strategy approach Because each strategy for demonstrating equivalence varies along a concrete to abstract continuum, it is implicit that researchers develop a multi strategy approach to proving cross-cultural equivalence in measurement. Although this might seem an onerous task, Hui and Triandis (1985) assure it is not: ... one can improve an instrument by proper translation techniques and then establish conceptuallfunctional equivalence as well as instrument (constructs as operationalized) equivalence by the nomological network method and by examination of internal structure congruence. After that, the response pattern method and regression methods can be used to test item equivalence and scalar equivalence. I mprovement can sometimes be made by modification of the format or contents of the instruments. (p. 149) What is important to remember is that single method strategies for improving or demonstrating equivalence are rarely adequate; rather, the researcher must 75 - --- ) prove equivalence at all levels of abstraction before precise quantitative generalizations can be made. 2.5.4 Assessment and Aboriginal North Americans Assessment of Aboriginal people can be problematic. On the one hand, '- communities and their members appear very much like North American culture, be it Canadian or American. On the other hand, however, Aboriginal cultures are distinct in many ways from the North American cultural milieu. As well, not all Aboriginal communities are alike and there are tribal or local differences made manifest by cultural diversity, language, environment, and acculturative influences. It is thus difficult to assume a particular stereotype for the concept "Indian culture" (Neeley & Shaughnessy, 1984). In discussing issues related to the assessment of Aboriginal North Americans and divergence from Euroamerican norms and from norms established within anyone Aboriginal group, a number of explanations have been offered for the apparent discrepancies. Ecological variations due to the impact of the environment on a particular cultural area in the form of deficit and developmental theories have been used to explain deviations in Aboriginal norms from Euroamerican norms. 2.5.4.1 Developmental change and assessment The culturally different patterns of raising children and the anxiety-laden learning conditions that some students encounter when entering a culturally unfamiliar environment, such as the school, have been offered as a developmental explanation for discrepancies between Indian and non-Indian children to exist in assessment data collected (McShane & Berry, 1988). As McShane and Berry explain: 76 _r- I .�,III. ------����;------__) The whole picture is one of a developmental pattern in which Indian children, reared in nurturant surroundings, enter schools that expose them to cultural discrepancies. Under culture-alien conditions they exhibit, by maturity, behavior that is perceived by the majority culture as either shy, withdrawn, and nonearning or, at the opposite pole of the flight-fight continuum, as rude, aggressive, and destructive to self and others. (p. 414) Certainly this developmental framework posited is one possible explanation that would take into account the manifestation and prevalence of extreme behaviours in Aboriginal child and adolescence populations. 2.5.4.2 Deficit and assessment Within the rubric of deficit and assessment are the concepts of genetic and physiological deficit. In terms of physiological deficit, it has been suggested that higher levels of alcohol abuse in some Aboriginal communities can lead to increased prevalence rates of Fetal Alcohol Syndrome (FAS) and its ensuing symptoms of hyperactivity and lower cognitive functioning (Report of the Steering Committee Medical Services Branch, 1991). Other physiological deficits which have higher prevalence rates in Aboriginal children and adolescents are otitis media and poor vision resulting in learning problems (McShane & Berry, 1988). Fink's (1990) research demonstrates social bonding, school failure, and susceptibility, defined as social awkwardness, low self esteem, and lack of social skills, as causing learning disabled students to be more at risk for delinquent behaviour and a negative psychosocial outcome than those students who are not identified as needing special education services. If the deficit argument posited is accepted, then a possible explanation could be offered for higher rates of delinquent behaviour and a negative psychosocial outcome in Aboriginal communities as a result of physiological factors. However, further research into this hypothesis would need to be undertaken before a definitive conclusion is reached. 77 � r I --- J Because of the differing perspectives offered, it is important that researchers ask themselves which hypothesis or combination thereof compliments the understanding and explanation of a given collected data base. Once this has been accomplished, then perhaps it will be possible to determine which factor accounts for the greater degree of variance within the population. Unfortunately, much of the research surrounding Aboriginal North American populations is circumspect and in order to further understand one's data, it will be necessary to examine common criticisms of cross-cultural studies and the ensuing design faults in research involving Aboriginal populations. 2.5.5 Testing and Aboriginal Students Over the years, a variety of tests have been administered to Aboriginal North American populations. Some of the tests used are: screening tests; intelligence tests; placement diagnostic tests; achievement tests; attitude tests; language ability tests; college entrance tests; reading tests; professional licensing tests; personality tests; and competency tests (Chavers & Locke, 1989). Because the literature is extensive and in some cases irrelevant, it will be necessary to focus only on those studies recently published in the field of personality, behaviour, attitude, and achievement testing. In a recent review of dissertations written on Indian education in the United States, Chavers and Locke (1989) reported that out of the twelve dissertations examined, seven were related to theory testing and explanatory research, while the others were related to "such things as applying the WISC-R to Native students" (p. 10). Although numerous studies question the applicability of the WISC-R in Aboriginal populations, it should be remembered that validating existing scales is more efficacious than the development of new tests and instrumentation (Kirmayer, 1994a). Much of the debate surrounding the testing 78 .' __ , ".��-"'"-. �------""I--�--- ; ..... of abilities, achievement, behaviour, and personality in Aboriginal populations is centered on what instrumentation is best suited for a particular task. However, before conclusions are drawn regarding the efficacy of instrumentation, studies need to be free of design faults. Unfortunately, a variety of researchers have admitted to severe structural faults in the testing of Aboriginal students (McShane & Berry, 1988). As a result, it is difficult to make recommendations and draw valid conclusions regarding a test's applicability. Recognizing some of the inherent difficulties in interpreting test results, the review of the literature will include recent studies in the field of assessment among Aboriginal populations. 2.5.5.1 Ability and achievement measurement in Aboriginal populations Research into ability and achievement testing among Aboriginal populations has been ongoing since the 1930's. Early studies conducted among Central Indians were undertaken by Telford in 1932, when he administered the Goodenough Intelligence Test to 225 Sioux pupils. Telford (cited in McShane & Berry, 1988) found mean scores of the students to be below that of American White and Black means. In a subsequent study using the same cohort conducted by Garth and Smith in 1937, it was reported that Indian children consistently achieved higher scores on the performance test than on the verbal test (cited in McShane & Berry, 1988). Similiar verbal and performance discrepancies have been confirmed by Sattler (1974) and later by McShane and Plas (1984a; 1988). In a review of the literature conducted by McShane and Plas (1984a) on the cognitive functioning of American Indian children, the authour's reported research with the Wechsler Intelligence Scale for Children (WISC), the Wechsler Intelligence Scale for Children-Revised edition (WISC-R), the Wechsler Adult Intelligence Scale (WAIS), and the Wechsler Preschool and 79 � I ___) Primary Scale of Intelligence (WPPSI; Wechsler, 1965b). It was found that all instruments supported the hypothesis that a discrepancy existed between verbal and performance IQ's for American Indian children (McShane & Plas, 1984a). Criticism of McShane and Plas's review of the literature has focused on the lack of critical evaluation of the theoretical and methodological problems in the studies cited (Brandt, 1984). Thus Brandt concluded: They [McShane & Plas] do not critically examine the methodological or theoretical problems in the studies they cite. This leads to two problems. They over generalize a few studies so that they appear to characterize all Native Americans, and they fail to draw obvious conclusions from other studies. They fail to consider the extreme heterogeneity of the Native American population. Their uncritical examination of sources leads them to build a profile for Native North American performance that does not control for English language fluency or for normal versus handicapped Indian populations. (p. 80) In response, M9Shane and Plas (1984b) argue the purpose of their study was not to endorse the WISC-R for use with Indian children. Rather, because the WISC-R is being used by educators, psychologists, and clinicians with Aboriginal children: how then can the instrument be better understood in light of the cultural change Aboriginal people are experiencing; and how better might their cognitive strengths and weaknesses be evaluated (McShane & Plas, 1984b)? Similiar findings regarding the Verbal-Performance IQ discrepancy among Aboriginal children has been corroborated in research conducted with the WISC-R (Goldstein, 1988; McShane & Plas, 1988; Seyfort, Spreen, & Lamer, 1980). Seyfort et at. reported findings which indicated a number of the items on the verbal subtest did not contribute significantly to the total test variance. In their research, Seyfort et at. attempted to restandardize the WISC-R for Aboriginal Indian children living on Vancouver Island based on the raw scores derived from the sample, N=177. The rationale behind this approach can be found in the explanation offered by Seyfort et at. that although an ethnic group 80 ------�� ,------ might be part of the larger standardization sample, interpretation of the scores for a particular group are not possible because intra group variance is submerged within the overall variance for the population. Furthermore, Seyfort et al. recognized that Indian children living on reserve have a cultural background which is at odds with the competitive, academic, and performance aspects of the WISC-A. As well, Seyfort et al, suggested the lower socio economic status and seasonal work characteristics of the Bands from which the sample was derived are not reflected on the verbal subtests of the WISC-R which are more geared towards white middle class values. This led Seyfort et al. to conclude: ... that clinicians using the WISC-R with Native children should interpret their findings with extreme caution. The norms developed from the present sample are based on small numbers and a limited geographical region (North Vancouver Island). They should therefore, be used with these factors and the previously discussed findings in mind. Because of the poor item validities reported, especially on verbal subtests, individual profile analysis as a tool of clinical inference should be avoided. (p. 23) What Seyfort et al, (1980) and other researchers using the WISC-R or other cognitive measures with Aboriginal children neglected to consider were the methodological recommendations offered for the testing of minorities (Berry, Poortinga et at. 1992; Irvine 1985; Irvine & Carroll, 1980; Massey, 1988 ). To address this inadequacy, Chrisjohn, Townson, Pace, and Peters (1987) conducted an internal and external analysis of the WISC-R with the Kainai in southern Alberta. Chrisjohn et aI., in structuring their methodology to reflect the research on the testing of minorities (Irvine; Irvine & Carroll), reported WISC-R results which are not at variance with the non-Native population. Because of the discrepancy with previous research findings, Chrisjohn et at. concluded that studies which do not follow recommended procedures for the testing of minorities are circumspect. In addition, Chrisjohn et al, stated their own study is 81 .,.", h: J-- inconclusive, as methodological deficiencies such as "small n's, collapsed across age groups" (p. 282), require the need for further research to enhance the study's validity and reliability. 2.5.5.2 Behaviour and personality measurement in Aboriginal populations The literature surrounding behaviour and personality assessment in Aboriginal populations is not as extensive as that found when discussing aptitude and abilities. Although behaviour and personality testing is not as controversial as the testing of abilities and aptitude in Aboriginal populations, it is at even greater risk for scientific opprobrium (Irvine & Carroll, 1980). Part of the problem lies in the extensive and varied research into personality and behaviour, resulting in various theoretical strands and subsequent interpretations. Given these variances, it is important to remember that the majority of instruments used and developed for the assessment of behaviour and personality are based on Western psychiatric nosology. Currently, the Diagnostic and Statistical Manual-4th Edition. (DSM IV; American Psychiatric Association, 1994) is the clinical diagnostic manual for practising psychologists and psychiatrists in North America. However, is it not true that the DSM IV is representative of Western psychiatric and psychological nosology and thus is culture bound in its' interpretation of distress and disorder (Kirmayer,. 1989)? This is a significant point when it is acknowledged that Western psychiatric and psychological nosology is based on the traditions of the American Psychiatric Association which have influenced the development of behavioural constructs in many of the behavioural and personality assessment instruments created. In attempting to understand the conflict which results when professional definitions of disorders, said to be objective and scientific, do not take into account the social context and are therefore, not culture dependent or 82 __) culture specific, Rubenstein and Perloff (1986) advise: ... in psychosocial work it is particularly important to recognize that there are many coexisting constructs of health and illness. Behavior deemed appropriate by some groups of lay-people may be defined as unhealthy by both professionals and by other groups of lay-people. Similarly healthy behavior, as professionally defined, may be seen as unhealthy by some nonprofessionals.This has led some researchers to claim that much of what is now called psychosocial disorder really reflects only culturally patterned variations of life style and social relations. The fact that health care providers see many of the psychosocial disorders as being most prevalent in poor and minority populations has reinforced this social and cultural interpretation, and led to the further argument that these phenomena are not really disorders. (pp. 321-322) McShane (1987) also recognizes tribal diversity as problematic in the development of culturally fair assessment instruments due to the variance in tribal conceptions of disorders such as depression. Because of the research involving the Minnesota Multiphasic Personality Inventory ( MMPI; Greene, 1978) and Aboriginal populations, it has been recommended that only with the development of local norms will assessment of behaviour and personality be clarified and equalized across responses (Butcher, Braswell, & Raney, 1983; Dana, 1986; Hoffmann, Dana, & Bolton, 1985). Since the role of acculturation in assessment is problematic, as it differs both between and within groups, awareness of cultural issues needs consideration in the interpretation of test scores (Butcher et al.; Hoffmann et al.). 2.5.6 Culturally Appropriate Assessment Practices In reviewing the literature as it relates to testing and assessment across cultures, Irvine and Carroll (1980) noted that methods of administering tests and the subsequent interpretations of tests scores were the guiding questions posited by researchers. Irvine and Carroll's work was a review of research since the 1960's and was subsequently used as a basis for the testing and assessment of minorities (Berry, Poortinga et aI., 1992; Chrisjohn et aI., 1987; 83 --�"I;;, ------�- ."J Irvine 1985; Massey, 1988). Recommendations articulated by Irvine (1973), and developed on in Irvine and Carroll, became the guiding principles for researchers as to the methods by which test conditions can be manipulated for I experimental purposes. The seven guiding principles outlined by Irvine (1973) are: 1. Every test response must be learned. No assumption should be made about a person's ability to respond in the manner required by the test. . Both test materials and methods of recording answers must be fully understood. 2. Often, in other cultures, the instruction to not turn back to previous tests is often bewildering and distressing to the people taking the test. Hence, omnibus tests booklets are undesirable. Each test should be separate from every other and should be accompanied by its own instructions. 3. Instructions should be oral, not written, since the ability to read instructions is not part of the test situation. Flexible, visual aids should demonstrate each type of test item and response, preferably by building up the components through the use of plastigraph or flannel graph techniques that allow cut-outs to be stuck to boards so that a large class can see. 4. If translations for test instructions are necessary, these should be literal. They should be idiomatic expressions of the intent of the test demonstrator, who ideally should be from the same ethnic group as those taking the test. This will reduce the possibility of extraneous motivational influences on test scores. 5. Supervised practice for each test is essential to make certain that the test instructions are understood. 6. Familiar test material should be given first, so that early attempts at recording answers becomes more automatic, unusual or abstract materials can then be presented. 7. The climate of testing must be convivial and dramatic and enjoyable as possible. Fully trained testers should employ every strategy to make the situation positively reinforcing and enjoyable. (p. 462-463) The rationale for adopting one or several of the principles outlined is for the determination, through multivariate statistical techniques, of whether or not the 84 .,"J '" "'I,:, ---r 1 ______��1,�i __ test measures what it proports to measure in a given culture (Irvine & Carroll). However, interpretation of the tests for clinical purposes is hazerdous, as test interpretation is based on standardization principles which specify administration protocol. Kirmayer (1994a), in reviewing mental health research on Aboriginal Canadians, attempted to resolve the emic-etic debate as it relates to behavioural and personality testing among Aboriginal populations. Kirmayer recommended that both emic, indigenous expressions of distress and disorder, and etic, heterogeneous standardized measures which allow for comparison in research, be incorporated in a given study. Although emic measures are more likely to be reliable and valid, they are specific to the culture or community in which they are developed and do not allow for comparisons to be made with other research findings (Kirmayer). Rather than attempt to develop instruments specific to a particular community, which is both time consuming, too difficult, intrusive, and idiosyncratic, it is recommended further research examine the modification and validation of scales used in culturally appropriate instruments (Kirmayer). In assessing culturally different exceptional children, Olion and Gillis-Olion, (1984) have noted that at times tests and outdated assessment practices have restricted the educational opportunities of children from different cultures. To rectify the problematic nature of assessment with culturally diverse children, it is suggested that assessment must involve both teachers and parents, be ongoing, multifaceted, and indicative of both the child's strengths and weaknesses to allow for culturally fair early detection (Olion & Gillis-Olion). By so doing, Olion and Gillis-Olion conclude that nondiscriminatory assessment practices are developed for the assessment of culturally diverse children. Sugai (1988) in discussing the educational and behavioural assessment of 85 the culturally diverse student endorses the use of reliable and effective assessment and evaluation practices. Rather than simply measure student performance in terms of strengths and weaknesses, Sugai recommended the student be assessed and/or evaluated in the context in which student performance or behaviour occurs. In order that the clinician be able to identify if indeed a problem exists, Sugai recommended the following guiding questions be asked before a problem is identified and special education services required: 1. Have several independent referrals been made? 2. How is the problem behavior operationalized or defined? 3. Is the behavior functionally different from some comparison or standard, for example, peer group? 4. Have there been dramatic changes in the individual's behavior in relatively short periods of time? 5. Have there been any significant life events in the student's or family's recent history? 6. Does the behavior interfere with the student's academic progress? Peer or adult relations? Community functioning? 7. Is the behavior destructive of property or injurious to other people? (p.70) Further to the issue of problem identification is the nature of the problem as defined by the communicative function of a behaviour and the concept of critical effect (Sugai, 1988). When Sugai discussed the communicative function of a behaviour, as it relates to the nature of a problem, he was referring to the "motivation and intent which drives a behaviour" (p. 79). For example, a child wishes to go to the washroom but is denied the opportunity by the teacher, the child in turn acts out and is asked to leave the classroom, the child then goes to 'the washroom. Conversely, the child can ask to work quietly outside of the 86 .�---�------�- classroom and then go to the washroom. The critical effect then can be the result of different forms of behaviour, and the clinician's responsibility in assessing the culturally diverse student is not only to ascertain behaviour but to examine both environmental and predisposing conditions to ameliorate biased referral and placement of culturally diverse students (Sugai). In developing an assessment model which is sensitive to the concept of cultural diversity, Massey (1988) described a tripartite multicultural assessment model. Massey went on to reiterate practices for test-based assessment which corroborates earlier work on cross cultural assessment and testing developed by Irvine (1973). Although Massey recognized that by 'testing the limits' and adapting tests to fit the cultural context in which they are delivered negates the standardized procedures developed for the instruments in question, he also acknowledged "that norm referenced interpretation is a largely unproductive endeavour among multicultural populations" (p. 27). In response to concerns surrounding culturally fair assessment practices, Massey developed a cross cultural assessment model, see Table 2.1. Table 2.1. A Cross Cultural Assessment Model 1: DATA COLLECTION 2: ASSESSMENT -Observation -Clinical interview/consultation with interpreter -Informal evaluation -Birthplace/age on arrivaillength of residency -Formal evaluation -Developmentallhealth history -Interm consultationlintervention -Educational history -Re-evaluation of skills -Family history -Ethnic affiliation -First language evaluation (informal) 3: INTERPRETATION & DIAGNOSIS -Cross-cultural factors influencing interpretation -Diagnostic formulationlholistic interpretation from a cross-cultural perspective -Intuition and experience-based interpretation -Processing strengths and weaknesses -Specific recommendations that fit the socio-cultural context Source: Massey, 1988; p. 23. 87 '111------ The model described by Massey (1988) is both a practical and simple method for the assessment of culturally diverse students. It is hierarchal in nature and allows for experiential, practical, problem solving of both the short and long term programming goals best suited to the student. The benefit of such a model is that "recommendations should include practical suggestions (rather than theoretical postulations) that can be implemented within the cultural context of the client's current environment" (Massey, p. 32). 2.5.7 Methodological Concerns Research in the field of testing and assessment, as it relates to Aboriginal people, is fraught with methodological concerns. Because of this, it is difficult to compare results from one study to the next. The process of research and data collection in many Aboriginal communities is impeded by a number of factors. According to Kirmayer (1994a), some of the factors which impede sampling in Aboriginal communities include: 1. Frequent travel out of the community into the bush. 2. Different perceptions of time and structures of daily routine making appointments unreliable. 3. Geographical distance and transportation access. 4. Traditional activities and religious cycles. 5. Caretaker responsibilities (especially for women). 6. Suspicion of outside researchers or burden of too much surveillance and previous research participation, resulting in low response rates. [7.] Random sampling ignores local community dynamics that may be crucial both to the conduct of a study and to its validity and translation into useful intervention. (p. 81) 88 � I Kirmayer goes on to state that in the collection of data the use of self-report questionnaires and standardized interviews for psychiatric diagnosis suffer from a number of limitations. Rather, focus groups and ethnographic methodologies like participant observation and key informants should be incorporated into studies to explore cultural idioms of distress (Kirmayer). Owing to the diversity of Aboriginal populations part of the difficulty in choosing a measure is that no one instrument is likely to work everywhere. Designing and validating new instruments is both time consuming and cost prohibitive and would be culture-specific to the group under study (Kirmayer, 1994a). Rather, according to Kirmayer: It [on how to collect data] argues in favour of using standardized instruments but addressing their limitations and possible sources of invalidity; for example, by adding culture-specific items, building in the possibility for using alternative diagnostic criteria of thresholds for caseness post hoc, emphasizing locally significant measures of disability and distress as outcome measures and combining epidemiological studies with parallel ethnographic research targeted at specific questions of validity and interpretation. (p. 72) 2.5.8 Summary In a review of the literature which encompassed research in medicine, psychology, sociology, anthropology, history, and education, a thorough analysis of the variables which place Aboriginal children and youth at-risk were explored. Topics covered included: social and economic demographics of Aboriginal people; the mental health of both Aboriginal and Canadian child and youth populations; the relationship between behaviour and intellectual functioning in children and adolescents; psychopathology across cultures; and cross-cultural assessment and instrumentation. Because of the literature review, a greater understanding of the variables which place Aboriginal children and adolescents at risk to later failure in life and at school can be developed. The debilitative effects of acculturation and the resultant psychological 89 change due to acculturative stress in Aboriginal communities was explored. The theory underpinning acculturative stress was corroborated by the social and economic demographic data and mental health indicators reported by Statistics Canada and the Medical Services Branch of Health and Welfare Canada. In this respect, the current study recognizes the rapid effects of acculturation which has significantly altered the ecological conditions among Aboriginal people. Poverty is common in many communities and its causes are the social inequalities and discrimination which face Aboriginal people today. Moreover, higher rates of unemployment and illiteracy, coupled with poorer living conditions, compound the ecological context in these same communities. The result of this rapidly shifting ecological context has seen the increased prevalence of children and families at-risk among many Aboriginal communities in Canada with these same children and families at-risk being denied many of the opportunities that mainstream Canadian society takes for granted. It is hoped that through the development of school-based interventions, partially based on the findings of the present study, the social, behavioural, emotional, and academic skills of children and youth at-risk will improve, allowing them to become equal participants in society. 2.6 The Present Study The present study explores the relationship between problem behaviours in the school based setting with the demographic, academic, and cognitive variables to determine the predictors of problem behaviours placing the child and youth population served by the Saskatoon Tribal Council at-risk. The initial screening phase of the Saskatoon Tribal Council study on exceptionality is the singular focus of this study. By not violating standardization principles during the initial screening phase, local norms can be created to assist researchers in 90 , .. ._-�------ determining cutoffs for the second phase of assessment in the Saskatoon Tribal Council study on exceptionality (Schwean & Saklofske, 1994). Owing to the large sample size and the extensiveness of the assessment battery, the data obtained lend themselves to a population study of the test scores and demographic data collected. Unfortunately, little research in the Canadian context exists in this domain among Canadian Aboriginal populations. As to why this might be, Kirmayer (1994a) suggests "opportunities have been lost because of the difficulties of preparing culturally appropriate instruments" (p.18). As the use of self-report measures in Aboriginal populations is circumspect and limited (Kirmayer, 1994a), teacher ratings of behavioural competence and academic performance are incorporated. In screening for cognitive development, a non-verbal test of reasoning ability is used. By employing standardized measures which partially alleviate the risk and opprobrium of epidemiological studies in Aboriginal populations, the present study conforms to some of the research recommendations for studies of Aboriginal people made in the literature review. In order not to compromise the validity of the study random sampling techniques were not employed, rather the total student population served by the Saskatoon Tribal Council and attending school during the period of data collection, comprises the subject sample. By not incorporating random sampling techniques the reliability and validity of the study is enhanced. However, the present study is limited in that it is not exhaustive in its presentation, incorporates only quantitative methodologies, and does not include an ethnographic or qualitative component. 91 ,___,.-- '"". --, -" ."""• .1 � ------",) 2.6.1 Research Questions 2.6.1.1 Research guestions #1 and #2 2.6.1.1.1 Rationale In the Ontario Child Health Study (Boyle et al., 1987; Offord et al., 1989, 1991; Sanford et al., 1992), teacher-identified impairment, in terms of school functioning, was highly related to teacher-identified disorder. Furthermore, studies have shown that students who manifest problem behaviours are at risk for failure in school and in life. These same students also score lower on intelligence tests and do not do as well academically. As well, research in the field of special education has found that students, identified as behaviourally disordered, emotionally disturbed, or learning disabled are more at risk for future delinquency than students who are not identified as special needs. For purposes of assessment and intervention, it is important to understand the relationship among the variables which can place children and youth at-risk. 2.6.1.1.2 Question #1 1. How important are academic performance as measured by the Academic Performance Rating Scale (APRS), cognitive ability as measured by the Matrix Analogies Test-Short Form (MAT-SF), age, sex, and school when each one alone is used to predict problem behaviours in a school based setting as measured by the Teacher Rating Scale of the Behavior Rating Profile-2nd Edition (BRP-2)? 2.6.1.1.3 Question #2 2. How important are academic performance as measured by the Academic Performance Rating Scale (APRS), cognitive ability as measured by the Matrix 92 XI ilFP#""",,_; .. __J4 J Analogies Test-Short Form (MAT-SF), age, sex, and school when they are used together to predict problem behaviours in a school based setting as measured by the Teacher Rating Scale of the Behavior Rating Profile-2nd Edition (BRP-2)? 2.6.1.2 Research question #3 2.6.1.2.1 Rationale In predicting problem behaviours in the general child and youth population, a number of factors have been found to correlate significantly. In terms of com mon psychiatric disorders, both sex and age have differential effects. Males tend to exhibit the more externalizing disorders like conduct disorder, aggression, and hyperactivity in the 12-16 year old age group, whereas females· tend towards the more internalizing disorders such as depression, anxiety, and social withdrawal in the 12-16 year old age group. It is also known, given the demographic trends reported for Canada's Aboriginal population, that the social and economic conditions facing on-reserve populations are significantly lower than the indicators reported for both off-reserve populations and the total Canadian population. 2.6.1.2.2 Question # 3 3. Is there a significant difference in problem behaviours in the child and youth population served by the Saskatoon Tribal Council based on age, sex, and school as measured by the Teacher Rating Scale of the Behavior Rating Profile-2nd Edition (BRP-2)? 93 r , . ....-\ I I 2.6.2 Assumptions 1. It is assumed the respondents to the demographic survey, the APRS, the TRS of the BRP-2, and the MAT-SF completed, to the best of their ability and honesty, the questions posited on the survey and instruments used in the Saskatoon Tribal Council special education research project. 2. It is assumed the sample selected is representative of the child and youth population served by the saskatoon Tribal Council. 2.6.3 Limitations 1. The study is limited in the instrumentation and demographic data collected as they may exclude areas of interest and are not exhaustive in their presentation. 2. Qualitative or ethnographic research methodologies were not incorporated addressing possible sources of invalidity in the study. 3. One school requested not to participate in the Saskatoon Tribal Council special education research. 2.7.4 Definitions Aboriginal: Refers to status and non-status Indians, Metis, and Inuit peoples. Academic Performance: Refers to academic achievement as based on teacher judgments measured by the teacher completed Academic Performance Rating Scale. At-Risk: Refers to the likelihood of failure at school or in life due to a number of factors that include economic, social, and physical variables which act as barriers to normal development. 94 -,01"''''' *"';';"" ...... ,.. , .'------"11_-...------ Bill C-31: Bill C-31 changed the meaning of the term "Status" and for the first time allowed for the reinstatement of Indians who had lost or were denied Status and/or Band membership in the past. The intent of Bill C-31 was to end many of the discriminatory provisions of the Indian Act, especially those which discriminated against Indian wom.en. First Nations Refers to Indian bands and their peoples in Canada, recognizing the historical reality of Indian people in that they have their own cultures, languages, traditions, and forms of government. Indian: Refers to registered or status Indians as defined in the Indian Act. The Indian Register is a list of all registered Indians which is kept by DIAND. Inuit: Refers to Native people, who have traditionally resided north of the tree line in the Northwest Territories, Labrador, and the northern coast of Quebec. Metis: Refers to those people who are of both Indian and non-Indian ancestry. Native: Refers to both status and non-status Indians. Non-Status Indians: Refers to those Indian people who are not registered under the Indian Act of Canada. Non-Verbal Reasoning Ability: Refers to estimates of the child's non-verbal reasoning ability and cognitive development as assessed by the student completed Matrix Analogies Test Short Form. 95 ,._11'* 'IQ .,...�.::, ---, �) On-Reserve: DIA'ND data include status Indians on-reserve and on Crown land and settlements, selected communities, and specified Census Divisions. Problem Behaviours: Are those student behaviours which occur in a school based setting as measured by the Teacher Rating Scale of the Behavior Rating Profile 2nd Edition. Status Indians: Refers to those people of I ndian ancestry who are registered under the Indian Act of Canada. 96 -r: I CHAPTER THREE METHODS AND PROCEDURES 3.0 Subject Sample The Saskatoon Tribal Council serves seven First Nations which include the three different linguistic groups Cree, Saulteaux, and Sioux. As well, each of the of seven First Nations served by the Saskatoon Tribal Council is autonomous unto itself. Within the Saskatoon Tribal Council, there are a number of communities ranging in size from 200 to 800 people. The subject sample included the students served by the Saskatoon Tribal Council, from Grades 1 to 12, who were registered and attended school 30 days prior to assessment. Only those students who live on-reserve are served by the Saskatoon Tribal Council, although they may attend off-reserve provincial and residential schools. Students from the Saskatoon Tribal Council attend over seventeen public schools, three band controlledschools, and three band controlled residential schools. Of the 809 students on the 1993 nominal role record, 289 attended schools located on the reserve. The rest attended provincial or residential schools located nearby. Those students who are registered in one of the seven bands that comprise the Saskatoon Tribal Council, but live off reserve, are not served by the Saskatoon Tribal Council and therefore, were not included as part of the subject sample. The apparent discrepancy between the sample population, 505 students, and the nominal role records can be accounted for as due to movement, such as change of residence both on and off reserve, and sporadic attendance 97 r, �� .. -- .J patterns. Of the 505 students assessed, 49% or 248 were girls and 51 % or 257 were boys. A frequency count of the sample by age is shown in Table 3.1. 3.1 Preliminary Protocol The initial screening was conducted by Ms. L. Robert, as part of the Saskatoon Tribal Council's study on exceptionality. Ms. L. Robert and Gordon Lobe, Superintendent of Education for the Tribal Council, made presentations to the various Bands served by the Saskatoon Tribal Council, informing them of the purpose and nature of the study. Band Council Resolutions (BCR) were passed by the Bands served by the Saskatoon Tribal Council supporting the collection of the data by granting permission to obtain assessment information from teachers, students, and parents. Parental consent was under the auspices of the individual Bands. In most cases parental consent was obtained in writing by the Bands, or when not possiole, verbally. Of the seven Bands served by the Saskatoon Tribal Council, six obtained consent through a consent form, while another obtained consent through a letter requesting parents of students to contact the Band office if consent was not forthcoming. Through this process, 13 parents of the students served by the Saskatoon Tribal Council did not provide consent, three were in disagreement, while ten could not be contacted. Those students whose parents did not provide consent were not included in the subject sample. For those students attending provincial schools, a similiar procedure was followed to obtain approval to conduct the study. Ms. L. Robert and G. Lobe directly contacted the provincial public school jurisdictions involved and made presentations to the District Boards of Education who served students from the Saskatoon Tribal Council. According to Schwean and Saklofske (1994), approval for data collection in those schools located in Prince Albert School 98 ______�� �T _ .----�------ Divisions was not as readily expedited. Although the majority of provincial educational institutions agreed to provide support to the Saskatoon Tribal Council, one school declined to participate. Table 3.1. Sample by Age Age Frequency Percentage of Total* 6 33 6.5 7 55 10.9 8 37 7.3 9 45 8.9 10 51 10.1 11 53 10.5 12 54 10.7 13 31 6.1 14 37 7.3 15 36 7.1 16 23 4.6 17 20 3.4 18 7 1.4 Other 23 4.6 Total 505 100 * percentages are approximations and may not total exactly one hundred percent. 3.2 Data Collection Data was collected during the period November 2, 1992 to June 30, 1993. Data collection was somewhat hindered due to "the number of different school jurisdictions involved and the latent mistrust by parents of anything related to special education" (Schwean & Saklofske, 1994, p.10). The initial screening phase consisted of the Academic Performance Rating Scale, the Teacher Rating Scale of the Behaviour Rating Protile-znd Edition, the Matrix Analogies Test-Short Form, and a short demographic questionnaire. Students were 99 ""\ I .. _- administered the MAT-SF according to test administration protocol dictated in the MAT-SF manual (Naglieri, 1985b). In this respect, the MAT-SF was administered in group to all students between Grades 2 to 12. For those students in Grade 1, the MAT-SF was administered individually or in groups of three. In most cases, the MAT-SF was administered during class, otherwise Ms. Robert administered the instrument when in-school times were not provided. In situations where the student moved and was not served by the Saskatoon Tribal Council, the case was culled from the sample. The teacher report form of the BRP-2 and the APRS were also completed for each student as part of the initial screening assessment. Individual cases were then matched with the teacher-completed demographic questionnaire, the BRP-2, and the APRS. The frequency count for the screening instruments are shown in Table 3.2, with incomplete and uncompleted instruments counted as missing. In total, data were collected on 505 students during the initial screening sample using the tests and rating scales described. 3.3 Instrumentation The instrumentation under discussion include the Matrix Analogies Test Short Form (MAT-SF), the Academic Performance Rating Scale (APRS), and the Behaviour Rating Profile-2nd Edition (BRP-2). The MAT-SF is a test of non verbal reasoning ability completed by the student and is delivered individually or in small groups, while the APRS and the Teacher Rating Scale (TRS) of the BRP-2 are based on teacher ratings of academic achievement and problem behaviours manifested in a school based setting, respectively. 100 r I "11----- Table 3.2. Sample by Instrument BRP-2 MAT-SF APRS Valid Cases 504 496 497 Missing Cases 1 a a Total 505 505 505 3.3.1 Teacher Ratings Teacher ratings of behaviour and academic performance have been proven to be of benefit when conducting psychological and behavioural assessments. This is so because teachers are able to observe students daily, in a relatively standardized school environment, and in relation to their peers, therefore, making the information obtained from behavioural rating scales methodical and systematic (Lupo & Forman, 1991; Sattler, 1992). Rating scales are also a fast, economical way of quantifying information for screening efforts, epidemiological studies, and program evaluations (Edelbrock, 1988). In the OCHS, which combined teacher, parent, and youth ratings, it was concluded that of the three informants, the teacher's perspective was the most uniform (Sanford et al., 1992). Moreover, it has been demonstrated that teachers "are better qualified to report on classroom behaviour, peer relations, academic performance, and behaviours such as difficulty following directions, inattentiveness, and ability to complete tasks" (Edelbrock, 1983, p. 296). In reviewing the literature on the psychometric properties of teacher ratings for behavioural, adaptive, and academic measures, it was found that teacher ratings are psychometrically adequate (Hoge, 1983). In discussing the relationship between teacher judgments and psychometric instruments in the identification of learning disabled and non-handicapped students, Gresham, Reschly, and Carey (1987) stated: 101 \ , ...... 1 One way of viewing teacher judgments is that they represent a summary or evaluative conclusions based on comprehensive and representative sample of behaviour observed in the classroom over a period of time. In terms of content validity and the domain sampling model, teacher judgments most certainly are based upon a much wider and comprehensive sampling of the content domain of achievement and classroom behaviour than standardized tests [of intelligence and academic performance]. As such, teacher judgments representatively sample the domain of interest in relation to school performance. (p. 544) Since the efficacy of teacher ratings and judgments of academic performance and classroom behaviour has been demonstrated, and because the present study relies on the quantification of data similiar in nature to epidemiological studies, the incorporation and reliance on teacher ratings of academic performance and problem behaviours in a school-based setting is both justified and warranted. 3.3.2 Behaviour Rating Profile The BRP-2 is composed of the same instruments as the Behaviour Rating Profile (BRP; Brown & Hammill, 1978) as developed by Brown and Hammill in 1978, 1983, and 1990. Although the item pools or administration of the instruments have not changed, subsequent editions were warranted to accommodate the inclusion of adolescents in the normative sample and to report ongoing validity and reliability studies. Brown and Hammill (1978, 1990) consider the BRP-2 to be an ecological assessment of behaviour and emotional problems from a variety of settings and perspectives. For the purposes of the Saskatoon Tribal Council study on exceptionality, the teacher completed report of student behaviours will receive primary attention as that was the BRP-2 instrument used as a behavioural screen. However, relevant discussions surrounding the psychometric properties of the complete BRP-2 instrument package will be considered. 102 � -r; , ;:, ...... ,.------�� --I The Teacher Rating Scale of the BRP-2 is composed of thirty items which describe problem behaviours in the school setting. Although, factor loadings have not been completed for the BRP-2 and specifically the TRS, it is generally agreed that the items used in the TRS are representative of teacher perceptions of inappropriate or problem behaviours in a school-based setting (Broughton, 1985; Ellers, Ellers, & Bradley-Johnson, 1989; Kratochwill, 1985; Witt, 1985). In fact, the authours of the BRP-2 suggest that one of its uses is "to help identify students who are believed to have emotional, behavioural, or personal and social adjustment problems" (Brown & Hammill, 1990, p. 5). The sentence stems on the TRS used to identify problem behaviours in a school based setting include item descriptors on verbal aggression, arguing with the teacher and classmates, disruptiveness, swearing, in school referrals for discipline, doing homework, following directions, motivation and interest, laziness, stealing, lying, bullying, cheating, passivity and withdrawal, overactivityand restlessness, nervous habits, daydreaming, lack of concentration, number of friends among classmates, unacceptable personal habits and hygiene, following directions, self-centredness, and academic achievement (Brown & Hammill, 1990). In rating the student, the teacher is asked to rate the student for each item on the basis of a four point descriptive Likert scale converted to a numerical weighting ranging in options from 0 (very much like the student) through to 4 (not at all like the student), resulting in a raw score which can then be converted to a percentile rank based on the norms provided in the BRP-2 examiner's manual. High scores on the BRP-2 indicate the absence of maladaptive behaviour, whereas low scores indicate a number of problem behaviours. In one of the first test reviews of the BRP, Kratchowill (1981) critisized the instrument because of its lack of sensitivity relating to the chroniCity and 103 .. ---�---- frequency of specific behaviours observed: Ostensibly, some teachers may have known the child for years and could rate behaviours occurring over varying periods of time. A problem such as "Disrupts the classroom" may be occurring for months while "Swears in class" may be a more recent problem. Thus the scale would not be sensitive to the discrimination of duration of those behaviours occurring over a several-month period of time versus those occurring within the relative recency of the past few weeks. (p.286) Due to this finding, Kratchowill (1985) recommends the clinician interpreting the instruments ask the respondent further information relating to the frequency and duration of a specific behaviour. Similarly, it is suggested the clinician determine the situation in which the behaviour occurs when planning intervention strategies (Kratchowill, 1985). Witt (1985) in his review of the BRP found wanting the description of specific behaviours as presented on the Teacher Rating Scale. According to Witt, "because of the lack of behavioural specificity, the meaning of an item could vary depending on the interpretation of the respondent" (p. 166). Nevertheless, the BRP has proven to be a stable screening instrument in the quantification of teacher's perception of classroom behaviour (Ellers et aI., 1989). But in Witt's opinion, this does preclude that the BRP-2 be the sole measure of behaviour. Rather Witt stated, "the BRP should presently be confined to use only as an adjunct measure in the identification of children who are behaviourally deviant" (p. 131). Given the restrictions outlined by Witt, it should be noted the purpose of the BRP-2 in the Saskatoon Tribal Council special education study was to act as a screen, in keeping with current research on the psychometric properties as they relate to the TRS of the BRP-2. Furthermore, the strength of the BRP-2Iies in its ability to assess across environments (Bacon, 1989; Posey, 1989) and to assist the clinician in planning programming involving the 'behaviourally disordered and learning disabled, whereas it weakness lies in discriminating behaviour among mentally retarded 104 , .�------ ...... ) and other population groups (Slate, 1983). In a study of learning disabled students, both regular and resource room teachers were asked to rate student behaviour. Although test-retest reliability coefficients for each rater proved high, .84 for regular teachers and .90 for resource room teachers (Lupo & Forman, 1985), the interrater reliability was low; "[the] resource room teachers rated the learning-disabled students' behaviour as more appropriate than did regular classroom teachers" (pp. 331). This result is not surprising considering the ecological context of behaviour in a specific situation and the subsequent interpretation of that behaviour among different raters. For the purposes of this study, only classroom teachers completed the TRS of the BRP-2. 3.3.3 Academic Performance Rating Scale The other teacher-completed rating scale used in the Saskatoon Tribal Council special education study was the Academic Performance Rating Scale, see Appendix 2. The purpose behind the development of the APRS was to provide an instrument which accurately assessed teachers' perceptions of overall academic performance and skills deficits associated with antisocial or disruptive behaviour which could also be used to monitor the efficacy of treatment programs (DuPaul & Rapport, 1991). As the present study sought to explore predictors of problem behaviours, an instrument which assesses the academic performance of children and youth at-risk is warranted. The APRS can be scored to yield a total score and scores for the subscales: Learning Ability, Academic Performance, Impulse Control, and Social Withdrawal (DPaul et aI., 1991). The advantage to using the APRS is that it provides information pertaining to the both the child's academic and behavioural functioning. Furthermore, the academic ability and performance of the child is not judged on specific content or curricula and is not subject to 105 � I -..J language bias in the assessment of academic ability as would be the possibility if a verbal test of intelligence was presented to a second language learner (Cummins, 1988; Jensen, 1980). Although the instrument is not long, 19 items, it proports to provide four subscale scores; however, no published study of a four factor solution for the APRS could be found. In an earlier study conducted on the APRS by DuPaul & Rapport (1991), a principal components analysis found the three factor solution academic sucess, academic productivity, and impulse control to exist. Subsequently, DuPaul & Rapport calculated the coefficient alphas: The results of these analysis demonstrated adequate internal consistencies for the Total APRS (.95), as well as for Academic Success (.94) and Academic Productivity (.94). The internal consistency of the impulse control subscale was weaker (.72).Subsequently, the total sample was randomly subdivided (Le., n=242 and 245, respectively) into two independent subsamples. Coefficient alphas were calculated for all APRS scores within each subsample with results nearly identical to the above obtained. (p. 289). In considering the efficacy of the APRS, it is recognized that it should be used as part of a multi modal assessment battery or as part of an initial screening assessment, especially if other measures of cognitive and behavioural deficits are being administered (DuPaul & Rapport). For the purposes of the present study, only the total score for the APRS was determined so as to provide an overall measure of academic performance not based on specific content or curricula. This measure was warranted because of the apparent ambiguity in the research literature surrounding the APRS as having a three or four factor solution. Although a few of the items on the APRS relate to behaviour, both of the externalizing and internalizing type, all other items refer to the student's academic performance over the past week. As the use of the APRS is limited as a screen of intelligence, a cultural free test of non-verbal reasoning ability not based on language is included in the present study. 106 , j 3.3.4 Matrix Analogies Test-Short Form The MAT-SF is a 34 item test, designed for use as a screening instrument in a group or individual setting (Naglieri, 1985b). The MAT-SF items are composed of abstract figural matrices which are a test of non-verbal reasoning ability. Non-verbal items employing geometric designs, like the MAT-SF, are culturally reduced. Because of this they are best used with young children and those who have difficulty in using the language on tests of verbal item groups (Jensen, 1980). Specifically, the MAT-SF is recommended for use with populations who have limited English language proficiency, have a communication disorder, or limited language development (Robinson, 1987). Naglieri (1985a) provided the following guidelines in the examiner's manual when trying to distinguish when to use the MAT-SF: Use the MAT-SF 1. When rapid screening of students is desired; 2. When group administration is possible; 3. When screening for an ability-achievement discrepancy (the Multilevel Academic Survey Test is required ... ); 4. When stanine, percentile, or age equivalent scores are sufficient. (p.3) In a study conducted by Naglieri (1986) examining the test-retest reliability of the MAT-SF in a large sample of grade five students, it was found that the MAT SF "test-retest reliability coefficient is similiar to those reported [in the MAT-SF] by Naglieri [for the WISC-R sub tests and the Peabody Picture Vocabulary Test Revised edition (PPVT-R; Dunn & Dunn, 1981)]" (p. 137). The purpose of the MAT-SF is to act as a screening instrument, recognizing that the scores should not be over generalized but rather interpreted as estimates of the child's non verbal reasoning ability (Naglieri, 1985b). In a review of the Matrix Analogies Test-Expanded Form (MAT-EF; Naglieri, 1985a), the instrument was critisized because it was found not to discriminate among children in the older age group who are of superior ability nor among those in the younger age group who are 107 ,.Ii "�';' ; �! fWS=;;=- -MV"II M'U!17i19"" of below average ability (McMorris, Rule, & Steinberg,1989). A similiar situation, although not as great, exists with the MAT-SF, which is composed of half as many items as the MAT-EF. The MAT-SF should not be the sole instrument used to identify students at risk of academic difficulty. Rather it is suggested by Naglieri (1985a) that the MAT-SF be used in conjunction with measures of academic achievement, either as a screening instrument or as a single measure in a complete test battery. Because other traditional measures of reasoning ability are verbal in nature and hence biased against students whose first language is not English, the advantage to using a progressive matrix instrument, like the MAT-SF, as a screen according to Jensen (1980): uris to assess] potential academic talent that might easily remain undetected by parents and teachers or by the more conventional culture loaded tests of scholastic aptitude". (p. 648) However, given Jensen's (1980) claim that matrix analogy tests which contain abstract reasoning items, like the MAT-SF, are culturally free, one would expect no significant difference between the norms developed in the MAT-SF based on an American sample and a sample drawn from Canadian schools. To explore the hypothesis that no significant difference between Canadian and American children's test scores on the MAT-SF would exist, Saklofske, Yackulic, Murray, and Naglieri (1991) conducted a study where the MAT-SF was administered to a sample comprised of children (N=660), aged 6 to 11 years, from five different schools in a midwestern Canadian city. It was found that some of the items on the MAT-SF were not functioning as expected based on age and total sample. Furthermore, the scores of children, in the 6 to 7 age group, were elevated (Saklofske et al.). Thus Saklofske et at. to conclude that: While the mean scores and variances were very similiar for the two. groups, Canadian children scored slightly higher at the 6 and 7 year age levels.These small but important finding reinforce the need to examine the generalizability of a test's norms, despite its satisfactory reliability and 108 validity estimates, when employing an instrument in a new or different context. (p. 56) In recognizing the MAT-SF for both its benefits and limitations with the score interpretation sensitive to recent research findings, it must be concluded that for the present study, the efficacy of the instrument is proven and therefore, its use warranted. Moreover, because the MAT-SF is used in conjunction with other screening instruments which include teacher ratings of academic performance and problem behaviours allows for some variation in source and mode of assessment. 3.4 Demographics The demographic data collected pertained to the students name, age, sex, and school (either provincial or Band) attended. The demographic data gathered from these questions were solicited from the home room teachers who also completed the teacher report forms of the BRP-2 and APRS. 3.5 Data Analysis Due to the varied number of research questions posited pertaining to predictors of problem behaviours, a single method of data analysis is inappropriate. As well, because the current study is descriptive and cross sectional in nature, it lends itself to a quantitative analysis. Because of the variables outlined, data analysis will employ a number of methods available through the SPSS suite of programs. In defining the variables for purposes of data analysis, the dependent variable and criterion variable are problem behaviours as measured by the TRS of the BRP-2. Thus the predictor or independent variables for purposes of data analysis are academic performance as measured by the APRS, non-verbal reasoning ability as measured by the MAT-SF, age group, sex, and school. 109 As well, the sample range is restricted, for purposes of data analysis, to those students who are between 6 and 18 years of age, to enhance the internal validity of the study which incorporated instrumentation designed for this age span. Furthermore, the sample is collapsed into three age groups representing elementary students from ages 6 to 9 years, middle years students from ages 10 to 13 years, and high school students between the ages of 14 to 18 years. These age groupings follow normal grade groupings found in most provincial and Band schools facilitating the analysis of data along these lines. As well, descriptive statistics are provided for all instruments and the demographic data collected relative to the research questions posited. In predicting problem behaviours in a school based setting, when the variables cited in question #1 are used alone, correlation matrices are calculated to determine the observed strength of the linear relationship between the variables identified. Subsequently, when the variables are used together to predict problem behaviours as cited in research question #2, the variables are entered into a multiple linear regression analysis to determine a model which can predict problem behaviours in a school based setting. Data transformation of the MAT-SF and the APRS scores occur (Norusis, 1993) so as to approximate a normal distribution, controlling for outlier variance thereby, facilitating a multiple linear regression analysis. Since the means and the variances are not known for the sample and must be estimated from the data, Norusis recommends that the Levene test of normality be used to determine if the hypothesis of normality can be accepted for the data transformations. Collinearity diagnostics report the degree of linear relationship between the variables used to predict problem behaviours. To determine if there are significant differences as cited in research question #3, based on age group, sex, and school for problem behaviours as measured by the teacher rating 110 , ...... 'Th'1 t I � ------�----� _ ._------ scale of the BRP-2, a 3 x 2 x 2 ANOVA table is used. If significant main effects and interaction effects are demonstrated, a 2 way ANOVA table is constructed to ascertain in what groupings significant differences occur. To confirm differences between group means, a univariate analysis of variance is conducted. 111 'if *" "'" '. \ 7 �: I CHAPTER 4 RESULTS 4.1 Overview In reviewing the literature, it was found that children who are at risk for a negative psychosocial developmental outcome manifest behaviours and emotions which have been variously described as problematic, inappropriate, extreme, delinquent, or disordered. The present study sought to explore the predictors of problem behaviours in the school based setting to determine which factors can assist in placing the child and youth population of the Saskatoon Tribal Council at risk for failure in school or life. As well, problem behaviours reported were compared between the groups', age, sex (male or female), and school (Band or provincial) to determine if the rating of problem behaviours varied as a function of the group, thereby, enhancing the reliability of the study for the population under consideration. Prior to conducting the data analysis, it was decided to limit the sample to those students who were between 6 to 18 years of age at time of assessment. Because the instruments used to measure problem behaviours, academic performance, and non-verbal reasoning ability were not standardized and norm referenced on populations under 6 years of age or over 18 years of age, the sample was marginally reduced. The sample was also grouped according to age into the 6 to 9 years of age, 10 to 13 years of age, and the 14 to 18 years of age groups. The decision to categorize the sample into the age groups outlined was modelled after the grade classification of children used in schools. Age groupings were then 112 .... -- I based on either Grades 1 to 3, Grades 4 to 8, or Grades 9 to 12.The rationale for the age grouping of students based on general grade level classifications was to enhance the analysis of the data to determine if a child's developmental level, as based on grade classification, effected the relationship between the variables in question. For the purposes of this study, one respondent who indicated attendance at a residential school was included under the category of Band school for further data analysis. This measure was warranted because the residential school in question is administrated by a Band which is part of the Saskatoon Tribal Council and therefore, should be considered as part of the sample which comprises the Band schools. Correlation coefficients were used to determine what variables, academic achievement, non-verbal reasoning ability, age (6 to 9 years of age, 10 to 13 years of age, and 14 to 18 years of age), sex (male or female), and school (Band or provincial), when used alone are related to problem behaviours in the school based setting. Subsequently, the variables were entered into a multiple linear regression analysis so as to build a model which can be used to predict problem behaviours in the school-based setting taking into account the effect of academic achievement, non-verbal reasoning ability, school (Band or provincial), sex (male or female), and age (6 to 9 years of age, 10 to 13 years of age, and 14 to 18 years of age). Analysis of variance was used to determine if there were any significant differences in problem behaviours between groups as based on school (Band or provincial), sex (male or female), and age (6 to 9 years of age, 10 to 13 years of age, and 14 to 18 years of age). 113 --.---=:::::==:,:",..="-=' •• ,::,.,..;;;..•••:...,...,. -'"�'II� .. --...... --�- -....' 4.2 Descriptive Statistics Table 4.1 provides the frequency counts of the number of students who completed the TRS of the BRP-2, the APRS, and the MAT-SF as based on age grouping. As is evidenced by the totals for the various instruments, the restricting of the sample to ages 6 to 18 years only marginally reduced the size of the sample from the total survey sample, N=505. As well, the fluctuation in sample size, as based on age groupings is relatively small at approximately ten percent. What is of interest in the frequency counts of the completed instruments as based on age groupings, is the lower numbers completed for all instruments in the 14-18 year age group. The discrepancy in the age group frequency counts is a result of the lower number of high school students on roll and served by the Saskatoon Tribal Council. This finding is not surprising given that over 50% of the Aboriginal people living on-reserve in Saskatchewan have a Grade 9 education or less, with only 1 .6% of the on-reserve population in Saskatchewan having received a high school diploma (The Social Trends Analysis Directorate, 1991). Table 4.1. Sample by Age and by Academic, Behavioural, and Cognitive Measure Age APRS BRP MAT-SF Freq. % Total* Freq. % Total* Freq. %Total* 6-9 170 35.9 170 35.5 158 34.0 10-13 186 39.3 189 39.5 188 40.4 14-18 117 24.7 120 25.1 119 25.6 Total 473 100 479 100 465 100 * percentages are approximations and may not total exactly one hundred percent. Table 4.2 provides the breakdown of the sample as based on school and sex by age grouping. The percentage difference, as based on frequency counts, 114 ._,_. "'. ,._,. 1 • , ..•••.• .. _------ I between sexes was greatest at seven percent in the 6 to 9 year age group. Otherwise the percentage difference between sexes was approximately five percent for the 10 to 13 year age group and two percent for the 14 to 18 year age group. In terms of school, for all age groups, a greater percentage of students attended provincial schools. This is especially true of the 14 to 18 year age group where over fifty percent more students attended a provincial school. For the total sample under consideranon, 50.8% are male and 49.2% are female, with 62.7% attending provincial schools and 37.3% of the students attending Band schools. Table 4.2. Age Group by School and Sex Age School Freq. %Total* Sex Freq. % Total* 6-9 Public 97 57.1 Male 91 53.5 Band 73 42.9 Female 79 46.5 Total 170 100 170 100 10-13 Public 112 59.3 Male 91 48.1 Band 77 40.7 Female 98 51.9 Total 189 100.0 189 100.0 14-18 Public 92 76.0 Male 62 51.2 Band 29 24.0 Female 59 48.8 Total 121 100 121 100 Total Sample Public 301 62.7 Male 244 50.8 Band 179 37.3 Female 236 49.2 Total 480 100 480 100 * percentages are approximations and may not total exactly one hundred percent. 4.3 Evaluating Assumptions of Normality Table 4.3 provides the distribution statistics of the instruments used for the total sample. In examining the distributions of the test scores, what is of interest is the low mean score, positive skew and negative kurtosis of the MAT-SF. As the standard scores reported for the MAT-SF in Table 4.3 are stanines, which in 115 .. --�---- a normal distribution would have a mean of 5 and a standard deviation of 2 (Naglieri, 1985b), it would appear that even though the bulk of the MAT-SF test scores are positively skewed and lie on the right side of the distribution, the mean score of the MAT-SF for the sample under consideration is below average. Moreover, because the distribution is platykurtic, indicates that it is more variable. Table 4.3. Distribution Statistics by Instrument for the Total Sample Mean Std. Dev. Kurtosis Skew BRP-2* 9.372 3.035 .025 -.247 APRS· 59.052 14.999 -.403 .024 MAT-SF** 3.683 1.759 -.290 .398 *standard scores;**stanines;·raw scores. Since the statistical procedures used in this study, correlation, multiple regression, and analysis of variance are sensitive to violations of normality and assume that all groups come from populations with homogeneous variances, it was necessary to determine what if any transformations were appropriate. One of the methods used to examine the assumption that the data come from a normal distribution is with normal probability plots (Norusis, 1993). SPSS allows for the plotting of both normal plots, which plot the deviations of the data points from a straight line and the detrended normal plot in which the points should cluster around a horizontal line in no particular pattern. In so doing, it was found that of the three measures used, the BRP-2 appeared not to violate the assumption of normality. This result is confirmed by the distribution statistics for the teacher rating scale of the BRP-2. Spread versus level plots were then used to determine the power of transformation for the independent variables academic performance and non verbal reasoning ability. A successful transformation will result in the points on a 116 -_...... '"' II' I� � I I .�------=� spread versus level plot clustering around a straight line resulting in a slope of zero. The power of transformation is calculated by subtracting the slope by one. Therefore, if a transformation is successful and the slope is zero, the power of transformation should equal one. Results of the transformations for the MAT-SF and the APRS are presented in Table 4.4. To determine if the sample had been derived from populations with equal variances, the Levene test for homogeneity of variances was used. The results of the data transformations for the independent variables has equalized the variances so that the spread versus level slope approximates zero. Because the Levene test for homogeneity of variances is not dependent on normal distributions, it is a particularly robust measurement for the purposes of analysis of variance and multiple regression (Norusis, 1993). Table 4.4. Results of Data Transformations Instrument Slope Levene Statistic df1 df2 Significance (Spread vs Level) APRS1 .016 .661 15 442 .8228 MAT-SF2 -.170 1.2002 15 442 .2678 Power of Transformation: 1. APRSe·671; 2. MAT-SFe1.672. 4.4 Research Question Results 4.4.1 Research Question #1 1. How important are academic performance as measured by the APRS, non verbal reasoning ability as measured by the MAT-SF, age group (6 to 9 years of age, 10 to 13 years of age, & 14 to 18 years of age) sex (male or female), and school (Band or provincial) when each one alone is used to predict problem behaviours as measured by the TRS of the BRP-2? 117 ·--�-- 4.4.2 Results and Discussion The correlations for the dependent variable, as measured by the TRS of the BRP-2, and the independent variables sex, age group, school, MAT-SF, and APRS are displayed in Table 4.5. Variables which are significantly correlated to the TRS of the BRP-2 include the APRS, MAT-SF, and sex. From observing the relationship between the variables, it is clear that academic performance, as measured by the APRS, is the strongest predictor of problem behaviours, as . measured by the TRS of the BRP-2, when used alone. Therefore, since r=.649 and r2 =.421, for the APRS, 42.1 % of the total variability is due to the correlation between academic performance and problem behaviours. Table 4.5. Correlations by Variable for Total Sample School Age Group Sex APRS MAT-SF School 1.000 Age Group -.144** 1.000 Sex .009 .022 1.000 BRP-2 -.063 -.050 .200** 1.000 APRS .064 -.093* .238** .649** 1.000 MAT-SF .035 -.097* .059 .219** .312** 1.000 *p<.05; **p<.01 However, other variables like the MAT-SF and sex are also significantly correlated to problem behaviours. For the MAT-SF where r=.219 and r2=.047, 4.72% of the total variability of problem behaviours, as measured by the TRS of the BRP-2, is due to the correlation with non-verbal reasoning ability, as measured by the MAT-SF. For the variable sex, where r=.200 and r2=.04, 4.00% of the total variability in problem behaviours, as measured by the TRS of the BRP-2, is due to the correlation with sex. A multiple regression analysis was also conducted as variables rarely act independently and multiple . influences can effect change measured in the dependent variable. 118 ... --'-.�-- 4.4.3 Research Question #2 2. How important is academic performance as measured by the APRS, non verbal reasoning ability as measured by the MAT-SF, age group (6 to 9 years of age, 10 to 13 years of age, and 14 to 18 years of age), sex (male, female), and school (Band, provincial) when they are used together to predict problem behaviours as measured by the TRS of the BRP-2? 4.4.4 Results and Discussion Table 4.6 presents the stepwise multiple regression data for the variables previously indicated. Since violations of assumptions require that data be transformed, the independent variables in Table 4.6 were transformed and entered. Stepwise selection was used because it is a combination of both forward and backward selection and determines the most parsimonious model. Appendix 3 contains the multiple regression results for the variables without any transformations taking place. In building a model which could predict problem behaviours, multiple regression analysis for the total sample, as well as for schools (Band and provincial), sex (male and female), and age group (6 to 9 years of age, 10 to 13 years of age, and 14 to 18 years of age) is conducted so as to enhance the reliability and internal validity of the study. To control the criterion by which variables are entered and removed from the equation, the F-to-enter (PIN) and F-to-remove (POUT) is set at 0.05 and 0.10, respectively. As would be expected from the correlation analysis, academic performance is the strongest predictor of problem behaviours, however, for the total sample, and the sample as based on age group and sex, school is also a significant predictor of problem behaviour when used in conjunction with academic performance. 119 -- " � --�--- ...._) Table 4.6. Stepwise Multiple Regression Predictions of Problem Behaviours (TRS of the BA P-2) Predictors Asq AsqCh F Sig. of F beta Total Sample N=480 Academic Performance .4636 390.727 .000* .6870 School .4777 .0141 206.256 .000* -.1188 Band Schools n=177 Academic Performance .5102 166.650 .000* .6767 Age Group .5271 .0169 88.623 .000* .1362 Non-Verbal Reasoning .5394 .0123 61.672 .000" .1212 Provincial Schools n=299 Academic Performance .4477 235.055 .000* .6717 Age Group .4556 .0079 120.915 .000* -.0978 Non-Verbal Reasoning .4643 .0009 83.201 .000* -.0974 Females n=233 Academic Performance .3961 146.239 .000* .6436 School .4161 .02 79.088 .000* -.1422 Malesn= 243 Academic Performance .4784 208.188 .000* .6925 School .4886 .0102 107.953 .000* -.1010 Age 6-9 years n=169 Academic Performance .3905 99.317 .000* .6408 School .4189 .0284 55.511 .000" -.1693 Age 10-13 years n=188 Academic Performance .5342 208.723 .000* .7315 ' School .5916 .0574 131.098 .000* -.2399 Age 14-18 years n=119 Academic Performance .4344 85.247 .000* .6658 School .4661 .0317 48.024 .000" .1783 *p<.01 120 , .. In interpreting multiple regression equations, R2indicates the degree of linear relationship between the independent and dependent variables. Thus for the total sample and the sample as based on sex and age group, academic performance accounts for the majority of the proportion of variation in problem behaviour as explained by the model. Furthermore, although school does meet the PIN requirements and contributes to the prediction of problem behaviours in the models developed, school's overall contribution is relatively low as evidenced by the R2 change and beta value. Nevertheless, since Band schools were coded as 0 and provincial schools were coded as 1, the negative beta value alludes to the greater predictive power of the equation in Band schools. When the sample is grouped according to either Band or provincial school, a somewhat different model is developed. Although academic performance is still . the strongest predictor of problem behaviours, it varies as a function of the school and accounts for 6% more predictive power in Band schools. Furthermore, because school is no longer a variable which is entered into the equation, the other independent variables, age group and non-verbal reasoning ability, pass the PIN criterion and are allowed entry using stepwise selection. However, the R2 change when age group and non-verbal reasoning ability are included in the model is minimal, which in conjunction with the beta values indicate the relative unimportance of these variables in predicting problem behaviours. As well, lack of collinearity between the independent variables is confirmed by the eigenvalues and variance proportions attributed to the eigenvalues for both Band and provincial schools. In this respect, Norusis (1993) states that: The variances of each of the regression coeffiCients, including the constant, can be decomposed into a sum of components associated with each of the eigenvalues. If a high proportion of the variance of two or more coefficients is associated with the same eigenvalue, there is evidence for a near dependency. (p. 357) 121 r -----�------ Table 4.7 presents the collinearity diagnostics for Band and provincial schools, with only those condition indexes which are elevated being reported. The collinearity diagnostics for the equations developed for the total sample, age group, and sex are not reported because the independent variables, school and academic performance, are highly independent. This relationship is similiar in nature to that for the total sample where, for the last eigenvalue, 99% or I almost 100% of the variance is accounted for by the constant and academic performance, while <1 % of the variance is attributable to type of school. In contrast, age group and academic performance are moderately related for both Band and provincial schools for the last eigenvalue in each equation as reported. Table 4.7. Collinearity Diagnostics Eigenvalue Condition Variance Progortions Index Constant APRS MAT-SF Age Group Band Schools .01196 17.444 .96727 .78240 .05339 .24808 Provincial Schools .01086 18.356 .96130 .83759 .01403 .16290 In determining collinearity, the researcher is looking for variables which have high proportions of variability associated with one eigenvalue (Norusis, 1993). Since the variance proportions for the variable academic performance is high in both equations, 73% and 82% for Band and provincial schools respectively, and the variance proportion is moderate for age group in both equations, 24% and 16% for Band and provincial schools respectively, while the proportion of variance is low for the MAT-SF in both equations, 5% and 1.4% for Band and provincial schools respectively, it would appear that 122 .�--. ------� ...... _) academic performance and age group are moderately collinear, while non verbal reasoning (MAT-SF) is independent to academic performance and age group. When examining the strength of the equations, as based on age group and sex, the predictive power of the equations is found to vary. The most striking difference is in the 10 to 13 year age group where 53.42% of the proportion of variability in predicting problem behaviours is accounted for by academic performance, and when school is added to the model, the predictive power of the equation increases to 59.16%. In contrast, the predictive power of the equation is lowest for the 6 to 9 year age group where 39.05% of the proportion of variability is accounted for by academic performance and increases slightly to 41.89% of the proportion of variability when school is added to the equation. For females, where 39.61% of the proportion of variability is accounted for by academic performance, the predictive power of the equation increases marginally by 2% when school is added to the equation. Similarly, in the 14 to 18 year age group, academic performance accounts for 43.44% of the proportion of variability when predicting problem behaviours and only increases slightly to 46.61% when school is entered in the equation. As well, when within group beta values for school are examined, it is found that only in the 14 to 18 year old age group does attending a provincial school, as opposed to a Band school, enhance the predictive power of the equation. Thus in predicting problem behaviours, the model developed is the most powerful in the 10 to 13 year age group, followed by the model developed for males only, and the total sample. In contrast, the predictive power of the equation is weakest in the 6 to 9 year age group and among females. In conducting multiple regression analysis, outliers are found to have an influence on the estimates of the parameters. SPSS can identify those outliers 123 .. .._- .�------= ...... 1 whose standardized residual values are greater than an absolute value of three. Table 4.8 provides data on the variables under consideration for the outliers identified. To enhance the predictive power of the equation, the outliers identified in Table 4.8 are not included in the stepwise multiple regression analysis. Considering the large sample size, the relatively small number of outliers identified for the total sample, enhances the goodness of fit for the model developed. Nonetheless, outliers are of importance because they indicate situations when the equation developed is not applicable. Thus when both academic performance and problem behaviours are high, or when behaviour is compotent and academic performance is very low, or when academic performance is above average and problem behaviours are negligible while non-verbal reasoning ability is low, a goodness of fit does not exist. Table 4.8. Outliers at 3 Standard Deviations Case Age Group Sex School APRS· BRP-2* MAT-SF* 42 6-9 years Female Provincial 64 63 35 234 14-18 years Female Band 57 1 38 432 10-13 years Male Band 41 75 4 438 14-18 years Male Provincial 16 75 75 ·APRS raw scores, p=59.05, std. dev.=15.00; *percentile ranks. 4.4.5 Research Question #3 3. Is there a significant difference in problem behaviours, as measured by the TRS of the BRP-2, between groups, as based on age group (6 to 9 years of age, 10 to 13 years of age, and 14 to 18 years of age), sex (male, female), and school (Band, provincial) in the child and youth population served by the Saskatoon Tribal Council? 124 4.4.6 Results and Discussion To determine the answer to the research question posited, an ANOVA 3 X 2 X 2 factorial design is constructed for the dependent variable measuring problem behaviours and the independent variables age group, sex, and school. Subsequently, if two-way or three-way interaction effects are present, ANOVA 2 X 2 factorial designs for the significant independent variables are created. In assessing problem behaviours, the TRS of the BRP-2 is the instrument used in the present study. To accommodate the interpretation of the ANOVA factorial designs for the TRS, the means and standard deviations of the variables age group, sex, and school, are provided in Table 4.9. Table 4.10 provides the results of the 3 X 2 X 2 ANOVA for the TRS. In examining the data, it is found that a main effect for sex F (1, 467) = 16.182, p<.01 and an interaction effect for age group by school F (2,467) = 3.351, p<.05 exist. To further tease out the main effect and determine if being male or female had similiar main effects and interaction effects, an ANOVA 3 X 2 factorial design for males and females with the independent variables age group and school is completed. Table 4.9. Descriptive Statistics for the Teacher Rating Scale of the BRP-2 Variable N Mean Std. Dev. Total Sample 479* 9.3176 3.0346 Age Group Age 6-9 years 170 9.4176 2.9726 Age 10-13 years 189 9.5873 2.9498 Age 14-18 years 120 8.9667 3.2331 Sex Male 244 8.7746 3.0541 Female 235 9.9915 2.8927 School Band 301 9.1236 3.4340 Provincial ·301 9.5183 2.7627 *480 cases were processed, 1 case, .2%, was missing. 125 .. --....------ Table 4.10. Analysis of Variance for Teacher Rating Scale of the BRP-2 by Age G roup, Sex, & School Source SS df MS F Sig. of F Main Effects 152.527 4 38.132 4.380 .002 Age Group 4.530 2 2.265 .260 .771 Sex 146.357 1 146.357 16.812 .000* School 6.593 1 6.593 .757 .385 2-Way Interaction 65.457 5 13.091 1.504 .187 Age Group X Sex 1.101 2 .551 .063 .939 Age Group X School 61.473 2 30.736 3.531 .030** Sex X School 3.235 1 3.235 .372 .542 3-Way Interaction 42.975 2 21.488 2.468 .086 Age Group X Sex X School 42.975 2 21.488 2.468 .086 Explained 336.373 11 30.579 3.513 .000 Residual 4065.481 467 8.706 Total 4401.854 479- 9.209 *p<.01; **p<.05; -480 cases were processed, 1 case, .2%, was missing. The ANOVA data run for the male sample is presented in Table 4.11. In examining the data, no significant main effects or interaction effects are present. However, when the analysis of variance is run for the female sample, there is a significant interaction effect for age group and school F(2, 229) = 5.822, p<.01, see Table 4.12. This indicates that the main effect and interaction effect witnessed in Table 4.10 is due to the elevated mean scores of the females which varies as a function of age group and school. 126 "'--�------ Table 4.11. Analysis of Variance for the Teacher Rating Scale of the BRP-2 (Males) by Age Group & School Source SS df MS F Sig. of F Main Effects 10.934 3 3.645 .390 .760 Age Group 4.700 2 2.350 .252 .778 School 9.455 1 9.455 1.012 .315 2-Way Interaction 10.990 2 5.495 .588 .556 Age Group X School 10.990 2 5.495 .588 .556 Explained 43.209 5 8.642 .925 .465 Residual 2223.393 238 9.342 Total 2266.602 243- 9.328 -244 cases were processed, 0 cases, 0.0%, were missing. Table 4.12. Analysis of Variance for the Teacher Rating Scale of the BRP-2 (Females) by Age Group and School Source SS df MS F Sig. of F Main Effects .910 3 .303 .038 .990 Age Group .802 2 .401 .050 .951 School .298 1 .298 .037 .847 2-Way Interaction 93.669 2 46.834 5.822 .003* Age Group X School 93.669 2 46.834 5.822 .003* Explained 115.895 5 23.179 2.882 .015 Residual 1842.088 229 8.044 Total 1957.983 234- 8.367 *p<.01; **p<.05; -236 cases were processed, 1 case, .4%, was missing Figure 4.1 graphically portrays that no significant interaction effect for males as based on age group and school is present. However, when females as based on age group and school is graphed, it is clear that a significant interaction effect is present. From observing Figure 4.2 it would appear that the significant interaction effect occurs in the 10 to13 year age group. 127 ...------ a Provincial Schools II Band Schools T I R S _ . 12 _ . m 11 e 10 _ . a n : � _--__�-::-_=-_:_:=::_-:__ :: ;_:�::�:�:::�::�_:�:: _: __ s _ . c 7 ; o 6�------�1------1------r 6-9 years 10-13 years 14-18 years e s I Age Group I Figure 4.1. Interaction Effect (Males) for Age Group and School a Provincial Schools II Band Schools T I R S 12 11 m e 10 a n 9 8 s c 7 0 6 r 6-9 years 10-13 years 14-18 years e s I Age Groupl Figure 4.2. Interaction Effect (Femal�s) for Age Group and School 128 .�--�------ To confirm that a significant effect occurs, a univariate post-hoc comparison between all groups as based on age group is made. The Scheffe' method is chosen because it is considered superior for simple between group comparisons when sample sizes are not the same (Evans, 1992). Significant effects, as based on age group, occur in both the provincial school and females attending provincial school groups. Thus the difference for the1 0 t013 year age group is significant with higher mean scores for the TRS of the BRP-2 in both the provincial school group F(2, 298)=6.417, p<.001 and the female provincial school group F(2, 144)=6.8065, p<.001. To further determine if type of school, Band or provincial, have similiar main effects and interaction effects for the TRS scores reported, an ANOVA 3 X 2 factorial design is run for each school grouping for the independent variable.s age group and sex. In Table 4.13, only the main effects for age group F (2, 295) = 5.477, p<.01 and sex F(1, 295) = 11.423, p<.01 are significant. When the analysis of variance is run for Band schools, using the same variables, the only significant main effect is for sex F (1,172) = 6.25, p<.05, see Table 4.14. This indicates that while sex and age group effect teacher ratings of problem behaviours in provincial schools, there is no interaction effect between the variables. In Band schools, only sex has a significant main effect and there is no interaction effect for age group and sex. 129 ...------ Table 4.13. Analysis of Variance for the Teacher Rating Scale of the BRP-2 (Provincial Schools) by Age Group & Sex Source SS df MS F Sig. of F Main Effects 170.751 3 56.917 8.019 .000 Age Group 77.756 2 38.878 5.477 .005* Sex 81.080 1 81.080 11.423 .001* 2-Way Interaction 21.099 2 10.549 1.486 .228 Age Group X Sex 21.099 2 10.549 1.486 .228 Explained 203.243 5 40.649 5.727 .000 Residual 2093.907 295 7.098 Total 2297.150 300- 7.657 *p<.01; '300 cases were processed, 0 cases, .0%, were missing. Table 4.14. Analysis of Variance for the Teacher Rating Scale of the BRP-2 (Band Schools) by Age Group & Sex Source SS df MS F Sig. of F Main Effects 91.337 3 30.446 2.656 .050 Age Group 15.718 2 7.859 .686 .505 Sex 71.741 1 71.741 6.259 .013** 2-Way Interaction 22.219 2 11.110 .969 .381 Age Group X Sex 22.219 2 11.110 .969 .381 Explained 115.707 5 23.141 2.019 .078 Residual 1971.574 172 11.463 Total 2087.281 177- 11.793 **p<.05; '178 cases were processed, 1 case, .6%, was missing. 4.5 Summary of Results The results indicate that academic performance and problem behaviours are moderately correlated. Sex and non-verbal reasoning ability are also significantly correlated with problem behaviours but the strength of the correlation is low. When problem behaviours are the criterion variable and non- 130 _...... r---" _ .. '--.=;" I \ verbal reasoning ability, academic performance, school, sex, and age group are the predictor variables, academic performance continues to be the strongest predictor, with type of school slightly increasing the predictive power of the multiple regression equation developed. The strength of the equation is strongest among the 10 t013 year age group and weakest among females. As would be expected, the multiple regression equation developed for Band and provincial schools is slightly different than the equation developed for the total sample and other groups. Although academic performance still remains the strongest predictor variable, other predictor variables entered into the multiple regression equation developed are age group and non-verbal reasoning ability. The strength of the equation developed for Band schools is slightly stronger than the one developed for provincial schools, with academic performance accounting for a greater degree of total variability. For all equations developed, except for the sample as based on Band and provincial schools, collinearity diagnostics indicate that the predictor variables are highly independent. In the sample as based on Band and provincial schools, a low but significant relationship exists between the variables academic performance and age group as predictors of problem behaviours. In determining significant differences for problem behaviours, analysis of variance results indicate that a main effect for sex and interaction effect tor age group by school exist. Furthermore, the results for the sample as based on males and females indicate that only in the female sample does a significant interaction between age group by school exist. In graphing the interaction effect, it appears that it is greatest among those females in the 10 t013 year age group either attending Band or provincial schools. Confirming this result, univariate analysis of variance report that in the 10 t013 year age female sample attending provincial schools, elevated mean scores for the TRS of the BRP-2 are present. 131 CHAPTER 5 DISCUSSION 5. 1 Overview of the Study The purpose of the present study was to investigate the predictors of problem behaviours which would place the student population served by the Saskatoon Tribal Council at risk for failure in school or in life. Specifically, the study examined predictors of problem behaviours as measured by the Teacher Rating Scale of the Behavior Rating Profile-2nd Edition with the variables sex, age group, type of school, non-verbal reasoning ability (as measured by the MAT-SF), and academic performance (as measured by the APRS). In developing a model to predict problem behaviours in the school-based setting, the correlates were analyzed both individually and together to determine which variables are the best predictors of problem behaviours. The present study also sought to explore if there was a significant difference in problem behaviours reported in the child and youth population served by the Saskatoon Tribal Council as based on age group, sex, and school. The sample derived for the present study included all students who were living on one of the seven reserves served by the Saskatoon Tribal Council and attending either a provincial, residential, or Band school. All schools serving students of the Saskatoon Tribal Council agreed to participate in the study except for one. Of the 809 students on the 1993 nominal role record, 289 attended Band schools witb the remainder attending provincial or residential schools. Those students who were registered in one of the seven bands that 132 ·�--�------=� comprise the Saskatoon Tribal Council but live off-reserve and were not served by the Saskatoon Tribal Council were not included in the subject sample. Thus the total sample population was 505 students with the discrepancy between the sample population and the nominal role records due to change of residence and sporadic attendance patterns. The subject sample was restricted to those students between the ages of 6 to 18 years to enhance the validity and reliability of the present study which incorporated instrumentation designed for this age span. As well, the subject sample was collapsed into three age groups representing elementary students from ages 6 to 9 years, middle years students from ages 10 to 13 years, and high school students between the ages of 14 to 18 years. Furthermore, the analysis of the data was conducted for the total sample and the sample as . based on sex, age group, and school. The purpose behind these groupings was fourfold. On the one hand, the division of the original sample into smaller samples enhances the reliability and validity of the present findings. Secondly, the research literature indicates that differential processes might affect the manifestation of problem behaviours, as based on age group. Thirdly, it would be expected that sex differences in the expression of problem behaviours would be present due to the design of the instrumentation used to assess problem behaviours. Finally, since not all students share the same educational context, as some attend Band schools whereas others attend provincial schools, problem behaviours might vary as a function of the school. Further explanation of the theoretical background supporting the groupings of the subject sample and the findings of the present study, with regards to the current research and future studies, will be elaborated on and discussed within the parameters of the present chapter. 133 I 1 ..�---- 5.2 Findings In determining independent predictors of problem behaviour for the total sample, it was found that academic performance, followed by non-verbal reasoning ability, and sex were significantly correlated, see Figure 5.1. Of the variables, academic performance accounted for over 42.1 % of the total variability of problem behaviours as measured by the TRS of the BRP-2. Although both non-verbal reasoning and sex are significantly correlated with Academic Performance Problem Behaviours r=.200 r=.238 Non-Verbal Reasoning Ability Figure 5.1. Correlation Model for Predictors of Problem Behaviours problem behaviours, the strength of the correlations is low. Non-verbal reasoning ability as measured by the MAT-SF accounted for 4.72% of the total variability of problem behaviour as measured by the TRS of the BRP-2 while sex accounted for only 4.00% of the total variability in problem behaviour as measured by the TRS of the BRP-2. This means that sex (p<.01, r=.200) and non-verbal reasoning ability (p<.01,r=.219) are poor predictors of problem behaviours in the school-based setting, while academic performance (p,.01, r=.649) is a moderate predictor of problem behaviours in the school-based 134 ...--, I setting. What is of importance for this finding is the relatively low correlation between non-verbal reasoning ability and problem behaviours. Rather it was found that a stronger relationship exists between non-verbal reasoning ability as measured by the MAT-SF and academic performance (p<.01, r=.312) as measured by the APRS, accounting for approximately 10% of the total variability in scores. Since numerous studies have found that students who are emotionally disturbed or behaviourally disordered score lower on intelligence tests and do not do as well academically (Farrington, 1991; Hallahan & Kaufmann, 1994; Werner & Smith, 1992; White et al., 1989; Worland et al., 1985; Toro et al., 1990; Ysseldyke & Algozzine, 1990), the present finding is of importance. In interpreting this finding it is important to realize that the MAT-SF is a non-verbal screen of intelligence, and since it is known that Aboriginal childrens' and youths' scores on intelligence tests are usually higher in the performance domain than in the verbal domain (Garth & Smith cited in McShane & Berry, 1988; Goldstein, 1988; McShane & Plas, 1984a, 1984b, 1988; Sattler 1974, 1992; Seyfort et aI., 1980), it is not expected that the MAT-SF, a non-verbal screen of reasoning ability, would be a strong predictor of problem behaviours. The present study partially bears this out as the strength of the correlation between non-verbal reasoning ability is low. However, in examining the mean score of the MAT-SF for the total sample, it is found that it is below average. In terms of the distribution of the MAT-SF, it is both positively skewed and platykurtic, resulting in a greater degree of variability in the distribution. Given these findings and the sample and locations from which the sample was drawn, the present study demonstrated that non-verbal reasoning ability as measured by the MAT-SF is a poor predictor of problem behaviours as measured by the TRS of the BRP-2. Instead, it was found that the strength of the relationship 135 between academic performance as measured by the APRS and non-verbal reasoning ability as measured by the MAT-SF was greater. In retrospect then, would a measure of verballQ have been a better predictor of problem behaviours given the relationship between academic performance, problem behaviours, and non-verbal reasoning ability? In multiple regression studies, using independent variables, sample sizes with over 200 subjects are preferred while studies incorporating less than a 100 subjects are not recommended (Kerlinger & Pedhazur, 1973). It is generally agreed that the larger the sample size the greater the power of the study and the more precise the statistical measurement ( Kerlinger & Pedhazur; Pedhazur & SChmelki�, 1991). In understanding why there is a slight difference in the equations developed for Band and provincial schools, it should be remembered that measurement error might account for some of the variance observed. This is so because the subject sample size for provincial schools (n=299) is larger than Band schools (n=177), by approximately 60%. Furthermore, even though students who attend both Band schools and provincial schools, for the purposes of the present study, live on reserve, they do not necessarily share a similiar educational context. Therefore, it would appear that even though academic performance is a stronger predictor of problem behaviours in Band schools, the strength of the equation developed for students attending provincial schools has greater power owing to the larger sample size for that group. As well, even though age group, to be discussed later, and non-verbal reasoning combine along with academic performance to enhance the goodness of fit for both the equations developed for Band schools and provincial schools, the overall contribution of the variable non-verbal reasoning ability is minimal. This finding indicates that further research into the functional equivalence of behaviour, ability, and achievement between settings needs to be investigated 136 ., .----�------ so that a valid, comparative, descriptive framework is developed (Berry, 1989; Kirmayer, 1994a; McShane & Berry, 1988; Rubenstein & Perloff, 1986; Sugai, 1988). In understanding why the equation developed for males has greater predictive power than the equation developed for females, it is important to consider that sex differences influence the expression of a disorder. In this respect, it is generally agreed that females tend towards the more internalizing disorders and manifest symptoms like depression, withdrawal, and somatization, whereas males tend to exhibit the more externalizing disorders like aggression and conduct disorder (Offord et aI., 1989, 1991; Patterson et aI., 1991). Even though the BRP-2 has not been factor analyzed, it is considered to have a single factor solution for conduct problems (Witt, 1985). Therefore, it is not unexpected that the TRS of the BRP-2 would under identify females who are internalizers but still at risk for failure in school. This finding clarifies why the strength of the equation developed to predict problem behaviours is stronger among males than females, as fewer items on the TRS of the BRP-2 identifying problem behaviours might have been endorsed for females who are internalizers and yet at-risk. According to research, differences in the strength of the multiple regression equation observed between age groups can be accounted for by the models which have been developed to interpret deviant or delinquent behaviour among children and youth (Farrington, 1991; Patterson et aI., 1991; Snyder & Huntley, 1990). As familial factors have been implicated as casual in the development and maintenance of aggression in children, it would be expected that instrumentation which is based on the rating of school-based problem behaviours would under identify children during their first few years of school. In understanding how item descriptors of a behavioural measure might effect the 137 c.. rating of problem behaviours, Farrington's research should be consulted. According to Farrington, descriptors of problem behaviours vary as a function of age with younger students 8 years of age being simply described as disobedient, while older students 12 and 14 years of age, are described as being not only disobedient but also quarrelsome, aggressive, rough at play, and resentful of criticism and punishment. Therefore, amongst younger children, teachers using the BRP-2 might have endorsed fewer items thus effecting the strength of the relationship between academic performance as measured by the APRS and problem behaviours as measured by the TRS of the BRP-2. Furthermore, because it is assumed, given normal intelligence, that academic achievement is determined primarily by time on task in the classroom and on homework assignments in the home, and since the child manifesting problem behaviours in the school environment spends less time on task than other children, it follows that he or she is at increased risk for academic failure (Patterson et al., 1991). As a result, the effects of academic failure cause dysphoric moods in the child, which are mediated by peer rejection, resulting in both peer rejection and academic failure (Patterson et aL).This explains why in the multiple regression equations developed that the predictive power of academic performance is strongest among the 10 t013 year age group and weakest among the 6 to 9 year age group, as the development of problem behaviours in the middle years leads directly to peer failure, poor academic performance, and then school failure. Conversely, in the 6 to 9 year age group, familial processes are deemed casual in the development and maintenance of antisocial behaviour (Farrington, 1991; Patterson et al.: Snyder & Huntley, 1990) which has not yet been transferred to the school and resulted in peer and school failure (Patterson et al.). Therefore, it would be expected that a model based on academic as a of behaviours in the performance predictor problem . 138 .. , .------ school setting, for the 6 to 9 year age group, is not as robust. As to why the predictive power of the equation is not strongest in the 14 t018 year age group can partially be accounted for by the high drop-out rates among Aboriginal and First Nation children (Report of the Steering Committee Medical Services Branch, 1991; The Social Trends Analysis Directorate, 1991), which subsequently indicates that the students most at risk for academic failure have already been failed by the educational system resulting in an equation in which academic performance has less predictive power. In determining if there were significant differences in problem behaviours between groups, it was found that males manifested more problem behaviours in a school-based setting than females. As well, when the sample was divided along the basis of sex to further tease out between group differences, it was determined that for the female sample, students in the 10 to 13 year age group were rated as having fewer problem behaviours in provincial schools than in Band schools. These findings confirm that the TRS of the BRP-2 identifies males as manifesting a greater number of problem behaviours in the school-based setting due to the construction of the instrument which is not as sensitive to internalizing disorders which are more prevalent among females. Even so, it must still be recognized that of those females who are identified as manifesting problem behaviours, the female students attending provincial schools in the 1 0 to 13 year age group were rated as having fewer problem behaviours than their counterparts attending Band schools. The significance of this finding underlies the possible difference in the educational context for First Nation students attending Band and provincial schools. 139 "j .�--�------� 5.3 Educational Implications The Education Act requires that all children and youth, of school age, attend an educational institution. Because of this, the school is the singular place where the multifaceted needs of children are dealt with on a daily basis. As well, it is generally agreed that preventive measures taken early in a child's life will reduce the risks, disadvantages, and future costs that the child, youth, or society may experience over time (Planning and Evaluation Branch, Saskatchewan Education, Training and Employment, 1994). Since the present study has found that academic performance is a strong predictor of problem behaviours, it follows that the school can be an ameliorative factor in the prevention and treatment of behavioural problems. As a result, if the child or youth at-risk is to receive effective remediation and become resilient to a negative psychosocial development outcome and not experience failure, then it is essential that the school act as an agent of change in the child's life. It is clear from the findings of the present study that there is a direct link between behavioural problems manifested in the school environment and academic performance. Research has also demonstrated that educational interventions need to develop the ability of children and youth at-risk to believe in their own capabilities to improve their academic performance and relations with others in the school (Connell et ai, 1994). Additionally, it has been found that once youth engage in behaviour which places them at-risk, there is a tendency for negative educational outcomes to occur (Connell et al.: Patterson et aI., 1991; Snyder & Huntley, 1990). In order for these destructive process to be reversed, it is important that those involved with children and youth at-risk, including both parents and educators, act to engage and Increase their academic involvement and thereby, reverse the negative processes which affect children and youth at-risk. .140 A' n.__�... �-.------�------ In terms of Canada's Aboriginal and First Nation populations, it has been reported that lower education levels and higher drop out rates than that for the total Canadian population exist (Report of the Steering Committee Medical Services Branch, 1991; The Social Trends Analysis Directorate, 1991). If First Nation students are to remain in school, it is important that the values reflected in the education system are not in conflict with the values of the family and community at large which the educational system serves (Planning and Evaluation Branch, Saskatchewan Education, Training and Employment, 1994). It is also necessary that the school involve both parents and the community so that the debilitative processes of acculturation are not reinforced in the educational system (Adams, 1989; Berry, 1980; Berry, Poortinga et aI., 1992; McShane & Berry, 1988). Moreover, it is generally recognized that strategies which focus on successful remediation are more successful when they build on the strengths of community identity and culture and are not imposed externally (Blum et aI., 1992). Therefore, if First Nation children are to succeed in school and not encounter failure and manifest problem behaviours, it is necessary that they have a strong sense of personal values which are not in conflict with the education system and are supported by their parents and community, who value school performance and are interested and involved in the education of their child (Planning and Evaluation Branch, Saskatchewan Education, Training and Employment). In this respect, the school should become a centre for community involvement which promotes parental and community participation in the education of their children at all levels. Because many of the problems which students face are not specific to the school and are the result of multiple influences which act on the child, it is important that service agencies develop a collaborative and consultative culture if resources available to the schools and educators are to be maximized 141 -.....,1 (Planning and Evaluation Branch, Saskatchewan Education, Training and Employment, 1994): Collaboration involving the adoption of cooperative principles, the establishment of partnerships and mutual goals, and the sharing of leadership, resources, and responsibilities is needed to address barriers to learning to enable children and youth to succeed in school and in life. In addition, the services provided by the different sectors are often fragmented and uncoordinated at the community level. Educators frequently do not know what resources are available or whom to call if an emergency arises with a student. In some locations services are not available. (p. 6) The need for the development of a collaborative culture among service agencies for children and youth at-risk is made even more imperative when it is realized that there will be a dramatic rise in the on-reserve school age population, not only because of higher retention rates in school, but also because of a continuing increase in the school age population (Hagey et aI., 1989). As a result, there will be an increased demand from Band run schools to further develop community education resources. Furthermore, Tribal Councils will continue to demand adequate resources from provincial and federal governments, as based on need, to address the growing demands of their school-age populations. Therefore, it is necessary that the service delivery of educational resources be developed in conjunction with the establishment of partnerships with Aboriginal and First Nation organizations whose mandate is to meet the essential needs of their children and youth. However, in order for educational service delivery to develop, it is important that both community and administrative stake holders share a common vision and direction in the development of educational opportunities for First Nation children and youth in Saskatchewan. It is not suprising then that policy initiatives advocate a cooperative, collaborative culture for the delivery of educational services (Planning and Evaluation Branch, Saskatchewan Education, Training and Employment). 142 .. " ",.",...,,--'\ 5.4 Research Implications 5.4.1 Culturally Sensitive Research The present study is a cross sectional analysis of the predictors of problem behaviours which place the student population of the Saskatoon Tribal Council at-risk. Because of this, the main emphasis of the present study was on the risk factors as determined by the instrumentation used and the demographic data collected. At present, little research in the Canadian context exists which explores the relationship between risk factors in Aboriginal populations. As to why this might be, Kirmayer (1994a) explains that research regarding the development of instrumentation which has been designed and standardized in Aboriginal populations has not occurred. As a result, researchers are wary of undertaking studies in which the reliability and validity of data gathering is in question. The research literature is clear in recommending that issues relating to functional equivalence and the emic-etic debate need to be addressed and resolved (Berry, 1989, Kirmayer). Even though emic measures of local idioms of distress, vocabularies, and categories of experience are likely to be more reliable and valid, they might not be comparable across populations. In order to prove that indeed psychometric equivalence exists and comparisons can be made cross-culturally using similar instrumentation, it is necessary to demonstrate or improve the various types of equivalence (Hui & Triandis, 1985). However, Kirmayer cautions that the revalidation and redevelopment of instruments specific to a given population is both time consuming and costly. Because of this, Kirmayer "argues in favour of working to modify and validate existing scales rather than starting from scratch each time" (p. 79). Leading Kirmayer to conclude: ... that all epidemiological research include an ethnographic component. This should use key informants and participant observation to explore cultural idioms of distress, as well as focus groups in which the proposed questionnaires are explored and critiqued.(p.79} 143 I"---=------=-�------ Even though local expressions of distress are likely to be unique and not comparable across populations, their incorporation will modify and/or validate ). standardized measures. Without such initiatives in research, studies will continue to be culturally laden and not valid or reliable interpretations of the populations under study. 5.4.2 Risk Factors and Resilience 5.4.2.1 Longitudinal studies As is the case in the present study, the majority of research concerning resilience and risk factors pertaining to children and youth is epidemiological in nature and therefore, occurring over the short term. Although the present study only concerned itself with predictors of problem behaviour which place the child and youth population at-risk and was limited to the variables under study, further research needs to explore the roots of resilience and the variables, factors, and mechanisms which place children at-risk and also allow them to overcome adversity (Werner & Smith, 1992). According to Rutter (1989), epidemiological studies which include a longitudinal component are receiving increased recognition because of the importance in recognizing why all individuals exposed to risk factors do not develop psychological disorders. Unfortunately, much of the research has focussed on adolescents and children in their middle years to the neglect of early childhood and longitudinal studies. Research which incorporates both quantitative and qualitative methodologies and monitors high risk populations over time to determine risk and protective factors including positive attributes and indicators of successful adaptation is both desirable and needed if social policy planning, interventions, and programming are to be effective. 144 .�------�------ 5.4.2.2 Developmental perspectives Rutter (1989) and Patterson et al. (1991) identify both continuities and discontinuities in the developmental patterns of psychopathology among children and youth. Etiology and age of onset are concerns which need to be investigated if at-risk and protective factors are to be understood. Patterson et al. has proposed two different models in the development of antisocial behaviour for children and adolescents, implicating familial processes in the early starter model and the delinquent peer group in the late starter model. Without a comprehensive understanding of the child developmental processes which are in place, opportunltles for remediation are lost. Rutter concluded that "a prospective epidemiologicaillongitudinal study starting in the preschool years and continuing into adolescence would be most helpful in delineating the processes involved" (p. 158). Other developmental niches identified by Werner and Smith (1992) which need to be examined in future epidemiologicaillongitudinal studies which focus on risk and resilience are the interface between the child and culture. Werner and Smith suggested physical and social settings, child care customs which include parental beliefs regarding a child's needs and wants, and sex role socialization need be investigated if protective mechanisms and risk factors are to be understood. In Aboriginal populations, it has been suggested that the culturally different patterns of raising children and the anxiety laden learning conditions that some students encounter when entering the school, a culturally unfamiliar environment, provide a developmental explanation for discrepancies in the assessment process to exist (Armstrong, 1993; McShane & Berry 1988). It is imperative then, that if further research into the at-risk and protective factors is to occur in First Nation populations, that issues pertaining to cross cultural psychopathology need be addressed. To facilitate this process, Kirmayer's 145 (1994a) guidelines to resolving the emic-etic debate and the reliability and validity of instrumentation scales should be incorporated if the study is not to be compromised. 5.4.2.3 Situation specificity The present study incorporated teacher ratings and a test of cognitive development. The teacher ratings were specific to the rating of problem behaviours and academic performance occurring in the environment of the school. In this respect, the reliability and validity of the study would have been enhanced if data was both multi modal and multi source. Although it is generally agreed that multi modal, multi source assessment is both desired and efficacious, little is known on how to integrate the differences as a function of the informant. According to the researchers involved in the OCHS (Boyle et aI., 1987), there is little empirical evidence regarding the integration of data from multiple perspectives. In the OCHS, it was found that not only were there disagreements between informants on individual cases, but that the correlates of the disorder also varied as a function of the informant (Offord et aI., 1989; Rutter, 1989). It is important then that psychological dysfunction continue to be measured along multiple dimensions and gathered from a number of informants and observers if the research is to remain culturally sensitive (Kirmayer, 1994a). 5.4.2.4 Multiple regression and replication Further to the issue of situation specificity is the need for future research to replicate the findings and thereby, enhance the reliability of the present study (Kerlinger & Pedhazur, 1973). This type of research is required as prevalence indicators have been found to vary among Aboriginal communities (Kirmayer, 146 1994a}, with the factors which place children and youth at-risk also varying as a function of the setting. Common criticisms of multiple regression analysis are the tendency of regression coefficients to change from sample to sample. It is generally conceded that findings which hold up using different subjects and conducted in different places can be trusted. However, studies which occur in only one local have reliability concerns. Certainly attempts were made, within the context of the present study, to determine if different groups within the sample had similiar predictors of problem behaviours. In replicating studies, it is not always necessary that the study duplicate methodologies and samples, rather minor changes to samples and variables can enhance the findings of previous research and broaden the knowledge base of a particular field of inquiry (Kerlinger & Pedhazur, 1973). Because of this, further research which replicates research studies, like the current one, in different places with different samples and with minor changes to the variables would enhance the reliability and internal validity of the present study, given similiar findings. 5.4.2.5 Gender bias The present study found that a significant main effect for sex in the rating of problem behaviours exists. Moreover, the mean score for males is about one half a standard deviation below that for females. Furthermore, there is a significant interaction effect among females in the 10 to 13 year age group. As well, it was found that females, 10 to 13 years of age, attending provincial schools were rated with fewer problem behaviours than their counterparts attending Band schools. These findings indicate that in terms of future research not only should behaviours which are externalizing in nature, like problem behaviours be identified, but also those behaviours which are internalizing in nature be identified. This measure is warrant�d because differential paths have 147 been developed for males and females both at risk and resilient to a disorder and susceptible to biological and psychosocial risk factors (Werner & Smith, 1992). By so doing, possible gender bias in the identification and prevalence of females who are at-risk will be mitigated. Future research also needs to examine the functional equivalence of problem behaviours between Band and provincial schools. Furthermore, the educational context experienced by students of both genders needs be determined to ascertain if situation specificity occurs in the expression and identification of problem behaviours. 5.5 Conclusion To assist children and youth at-risk, it is necessary to examine those processes which are in place to help determine how positive change can be implemented and bring about adjustments leading to successful adaptation. In terms of prevention and intervention in a school-based setting, the present study found that the facilitation of academic success continues to be one of the most important variables which can affect the manifestation of problem behaviours leading to subsequent academic or school failure. The responsibility of educators then is to reverse this negative transactional process. For children and youth to to succeed academically and to reach their potential, it is necessary that they encounter responsible adults who are caring and trusting allowing them to feel safe and secure within the context of the educational setting. The significance of the school in both community development issues and as an advocate of the values and beliefs of the community it serves is made manifest when the historical context in which education occurred among Aboriginal populations is recognized. The development of intervention programs should be directed towards all age groups. Moreover, the importance of intervention is made even more acute 148 ..".,....." .... i" • w�--�------�------ when it is realized that females at-risk might be under identified in the present study. Furthermore, of the females who are identified, it is found that those 10 to 13 years of age attending Band schools are more at risk for academic and school failure than their counterparts attending provincial schools. Given this finding, it is necessary that federal and provincial funding agencies continue to make available resources to Tribal Councils for the development of appropriate educational interventions if those children who are at-risk are to receive remediation. As well, the current delivery of educational services needs to facilitate a collaborative culture whereby interagency cooperation between federal, provincial, and First Nation agencies is accommodated. Without such a service delivery mechanism in place, agencies representing both federal and provincial jurisdictions will continue to offer fragmented and opposing services to First Nation children and youth. At present, within the Canadian context, there is a lack of valid and. reliable epidem iological research among First Nation populations. There are currently no published studies examining the longitudinal protective factors of resilient children and youth which allow them to develop successful adaptations to later adult outcomes. Assessment and diagnosis, whether it be preventive or ameliorative, needs to examine not only risk factors but those processes which are protective and allow the child or youth to become resilient to a negative developmental outcome. Future research also needs to be validated culturally and demonstrate the incorporation of both emic and etic methodologies if findings are to be comparable and studies are not compromised by weak research designs. Without such research initiatives, policy and programming initiatives will continue to be lacking as they fail to take into account the dynamics and developmental processes which effect First Nation children and youth. 149 REFERENCES Adams, H. (1989) Prison of grass: Canada from a Native point of view. Saskatoon, SK: Fifth House. American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders. (4th. �d.). Washington, DC: Author. Armstrong, H. (1993). Depression in Canadian Native Indians. In P. Cappeliez & R.J. Flynn (Eds.), Depression and the social environment (pp. 218- 234). Montreal: McGill-Queen's University Press. Bacon, E.H. (1989). Review of the Behaviour Rating Profile. In J.C. Conoley & J.J. Kramer (Eds.) The tenth mental measurements yearbook. (pp. 84-86). Lincoln, NE: The University of Nebraska Press. Beiser, M. & Attneave, C.L. (1982). Mental disorders among Native American children: Rates and risk periods for entering treatment. American Journal of Psychiatry, 139, 193-198. Berry, J.W. (1980). Social and cultural change. In HC. Triandis & R.W. Brislin (Eds.), Handbook of cross-cultural psychology: Social psychology. 5, (pp. 211-279). Boston, MA: Allyn & Bacon. Berry, J.W. (1989). Imposed etics-emics-derived etics: The operationalization of a compelling idea. International Journal of Psychology. 24, 721-735. Berry, J.W. & Bennett, J.A. (1992). Cree conceptions of cognitive competence. International Journal of Psychology, 27, 73-88. Berry, J.W., Kim, U., Minde, T., & Mok, D. (1987). Comparative studies in acculturative stress. International Migration Review, 21, 491-511. Berry, J.W., Poortinga, HY., Segall, M.H, & Dansen, P.R. (1992). Cross cultural psychology: Research and applications. London, ENG: Cambridge University Press. 150 _Jt' "!!:lI:Ii...'; � .� -:;:r �i <� - �� Berry, J.W., Wintrob, R.M., Sindell, P.S. & Mawhinney, T.A. (1982). Psychological adaptation to culture change among the James Bay Cree. In R. Rath, H.S. Asthana, D. Sinha & J.B.P. Sinha (Eds.), Diversity and unity in cross cultural psychology (pp.157-170). Alblasserdam, DU: Swets & Zeitlinger B.V .. Blum, R.W., Harmon, B., Harris, L., Bergeisen, L. & Resnick, M.D. (1992). American Indian-Alaska Native youth health. Journal of the American Medical Association, 267, 1637-1644. Boyle, M.H, Offord, D.R., Hofmann, H, Catlin, G.P., Byles, J.A., Cadman, D.T., Crawford, J.W., Links, P.L., Rae-Grant, N.L, & Szatmari, P. (1987). Ontario Child Health Study: Methodology. Archives of General Psychiatrv,44, 826-831 . ... Brandt, E.A. (1984). The cognitive functioning of American Indian children: A critique of McShane & Plas. School Psychology Review, 13,74-82. Brody, H (1987). The living arctic. London, ENG: Faber and Faber Limited. Broughton, S.F. (1985). Behaviour Rating Profile. In D.J. Keyser & R.C. Sweetland (Eds.), Test Critiques, IV (pp, 92-102). Kansas City, MO: Test Corporation of America. Brown, L. & Hammill, D.O. (1990). Behaviour Rating Profile: Examiner's manual. (2nd ed.). Austin, TX: PRO-ED. Brown, L. & Hammill, D.O. (1978). Behaviour Rating Profile. Austin,TX: PRO-ED. Butcher, J.N., Braswell, L., & Raney. D. (1983). A cross-cultural comparison of American Indian, Black and White inpatients on the MMPI and presenting symptoms. Journal of Consulting and Clinical Psychology, 51, 587-594. Chavers, D. & Locke, P. (1989). The effects of testing on Native Americans. Albuquerque, NM: Native American Scholarship Fund. (ERIC Document Reproduction Service No. ED 336445). 151 "" -:1" �. I __�__ � I ._------ Chrisjohn, R.D., Townson, S.M., Pace, D.F., & Peters, M. (1987). The WISC-R in a Native application: Internal and external analysis. In J. W. Berry & R.C. Annis (Eds.) Ethnic psychology: Research and practice with immigrants, refugees. Native peoples, ethnic groups. and sojourners. Amsterdam, DU: Swets & Zeitlinger. Connell, J.P., Spencer, M.B., & Aber, J.L. (1994). Educational risk and resilience in African-American youth: Context, self, action, and outcomes in school. Child Development 65, 493-506. Cronbach, L.J. & Meehl, P.E. (1955). Construct validity in psychological tests. Psychological Bulletin. 52" 281-302. Cummins, J. (1988). Empowering minority students. Ottawa, aNT: Social Services and Humanities Research Council. Dana, R.H. (1986). Personality assessment and Native Americans. Journal of Personalitv Assessment 50, 480-500. DuPaul, G.J. & Rapport, M.D. (1991). Teacher ratings of academic skills: The development of the Academic Performance Rating Scale. School Psychology Review. 20,284-300. DuPaul, G.J., Rapport, M.D., & Perriello, L.M. (1991). The Academic Performance Rating Scale. In R.A. Barkley (Ed). Attention deficit hyperactivity disorder: A clinical workbook (pp.64-65). New York, NY: Guilford Press. Driver, H.E., & Massey, W.C. (1957). Comparative studies of North American Indians. Philadelphia: American Philosophical Society. Dumaret, A. (1985). la, scholastic performance, and behaviour of sibs raised in contrasting environments. Journal of Child Psychology and Psychiatry. 26, 553-575. Dunn, L.M. & Dunn, L.M. (1981). The Peabody Picture Vocabularv Test Revised. Circle Pines, MN: American Guidance Service. 152 Eccles, W.J. (1988). The Canadian frontier 1534-1760. Albuquerque, NM: University of New Mexico Press. Edelbrock, C. (1983). Problems and issues in using rating scales to assess child personality and psychopathology. School Psychology Review. 12, 293-299. Edelbrock, C. (1988). Informant reports. In E.S. Shapiro & T.R. Kratchowill (Eds.), Behavioural assessment in schools: Conceptual foundations and practical applications (pp. 351-383)). New York: Guilford Press. Ellers, R.A., E"ers, S.L., & Bradley-Johnson, S. (1989). Stability reliability of the Behaviour Rating Profile. Journal of School Psychology. 27. 257-263. Evans, A.N. (1992). Using statistics in the behavioral sciences. (2nd ed.) . Scarborough, ONT: Prentice-Hall. Farrington, D.P. (1991). Childhood aggression and adult violence: Early precursors and later-life outcomes. In D.J. Peplar & K.H. Rubin (Eds.), The development and treatment of childhood aggression (pp. 5-29). Hillsdale, N.J.: Lawrence Erlbaum Publishers. Fink, C.M. (1990). Special education students at risk: A comparative study of delinquency. In P.E. Leone (Ed.), Understanding troubled and troubling youth (pp. 38-81). Newbury Park, CA: Sage Publications. Frymier, J. & Gansneder, B. (1989). Building a community for learning: Integrated school-based services. Regina, SK.: Saskatchewan School Trustees Association Research Centre. Frontera-Benvenutti, R.L. (1991). A comparison of two independent Spanish translations for the student questionnaire of the Behaviour Assessment System for Children (BASC). (Doctoral dissertation, University of Texas, (1991) Dissertation Abstracts International, 52 (9-B), 5014-B. Goldstein, C.R. (1988). Cautions concerning assessment of Native Indian students. Canadian Journal of School Psychology. 4, 17-20. 153 .------�-�------ Goldstein, H.R. (1986). Conduct problems, parental supervision, and cognitive development of 12-17 year olds. Psychological Reports. 59, 651-658. Graham, A. (1987) An overview of socio-demographic conditions of registered Indians living off-reserve, Ottawa, ONT: Health and Welfare Canada. Greene, L.L. (1978). The MMPI: An interpretive manual. New York, NY: Grune & Stratton. Gresham, F.M., Reschly, D.J., & Carey, M.P. (1987). Teacher as 'Tests": Classification accuracy and concurrent validation in identification of learning disabled children. School Psychology Review. 16, 543-553. Hagey, J.N., Larocque, G. & McBride, C. (1989a). Highlights of aboriginal conditions 1981-2001: Part I demographic trends. Ottawa, ONT: Department of Indian Affairs and Northern Development. Hagey, J.N., Larocque, G. & McBride, C. (1989b). Highlights of aboriginal conditions 1981-2001: Part II economic conditions. Ottawa, ONT: Department of Indian Affairs and Northern Development. Hallahan, D.P. & Kaufman, J.M. (1994). Exceptional children: Introduction to special education (6th ed.). Boston, MA: Allyn and Bacon. Health and Welfare Canada Medical Services Branch (1992). Aboriginal health in Canada. Ottawa, ONT: Health and Welfare Canada. Health and Welfare Canada Medical Services Branch Saskatchewan Region (1989). The health of the registered Indian population in Saskatchewan. Ottawa, ONT: Health and Welfare Canada. Hoffman, T., Dana, R.H., & Bolton, B. (1985). Measured acculturation and MMPI-168 performance on Native American adults. Journal of Cross-Cultural Psychology. 1 6, 243-256. 154 ��.,; ---'� .�------�------ Hoge, A.D. (1983). Psychometric properties of teacher-judgement measures of pupil attitudes, classroom behaviours, and achievement levels. The Journal of Special Education. 17, 401-429. Hui, C.H & Triandis, H.C. (1985). Measurement in cross-cultural psychology: A review and comparison of strategies. Journal of Cross-Cultural Psychology,.1§., 131-152. Indian Health Service (1992). Trends in Indian health. Rockville, MD: U.S. Department of Health Services, Public Health Service, Indian Health Service. INED (1973). Travaux et documents No. 64. Enquete nationale sur Ie niveau intellectual des enfants d'age scolaire. Paris, FR: PUF. Irvine, S.H. (1973) Tests as inadvertent sources of discrimination in personnel decisions. In P. Watson (Ed.), Psychology and race. London, ENG: Penguin Books, (pp. 453-466). Irvine, S.H (1985). What does research have to say abo minorities? In A. Samunda & A. In Wolfgang (Eds.), ____te=r-=c=ul=tu=r�"...... -...... _...-...... = assessment: Global perspectives (pp. 165-176).Toronto: C.J. Irvine, S.H & Carroll, W.K. (1980). Testing and assessment across cultures; Issues in methodology and theory. In HC. Triandis & J.W. Berry (Eds.), Handbook of cross-cultural psychology: Methodology. 2 (pp. 181-244). Boston, MA: Allyn & Bacon. Jastak, J.F., & Jastak, S.A. (1978). The Wide Range Achievement Test manual of instructions. (Rev. ed.). Wilmington, DE: Jastak Associates. Jensen, A.A. (1980). Bias in mental testing. New York, NY: The Free Press. Kauffman, J.M. & Hallahan, D.P. (1994). Exceptional children: Introduction to special education (6th ed.). Boston, MA: Allyn and Bacon. 155 �lf4 ,po ""'J� , Kaufmann, J.M. & Needler, RD. (1981). Behavioural disorders. In J.F. Kaufmann & D.P. Hallahan (Eds.), Handbook of special education. Englewood Cliffs, NJ: Prentice-Hall. Kerlinger, F.N. & Pedhazur, E.J. (1973). Multiple regression in behavioural research. New York, NY: Holt, Rinehart, and Winston. Kirmayer, L.J. (1989). Cultural variations in the response to psychiatric disorders and emotional distress. Social Science Medicine. 3, 327-339. Kirmayer, L.J. (1994a). Emerging trends in research on mental health among Canadian Aboriginal peoples (Culture and Mental Health Research Unit Report No.2 prepared for the Royal Commission on Aboriginal people). Montreal, QUE: Culture and Mental Health Research Unit, Institute of Community and Family Psychiatry, Sir Mortimer B. Davis-Jewish General Hospital & Division of Social and Transcultural Psychiatry, Department of Psychiatry, McGill University. Kirmayer, L.J. (1994b). Suicide in Canadian Aboriginal populations: Emerging trends in research and intervention (Culture and Mental Health Research Unit Report No.1 prepared for the Royal Commission on Aboriginal people). Montreal, QUE: Culture and Mental Health Research Unit, Institute of Community and Family Psychiatry, Sir Mortimer B. Davis-Jewish General Hospital & Division of Social and Transcultural Psychiatry, Department of Psychiatry, McGill University. Kniveton, B.H. (1987). Misbehaving peer models in the classroom: An investigation of the effects of social class and intelligence. Educational Studies . . .u, 161-168. Kratchowill, T.R (1981). Review of the Behaviour Rating Profile: An ecological approach to behavioural assessment. Journal of School Psychology. ll,283-288. Kratochwill, T.R (1985). Review of the Behaviour Rating Pronle, In J.V. Mitchell Jr. (Ed.) The ninth mental measurements yearbook (pp. 163-166). Lincoln NE: The Nebraska Press. ... University of 156 .�" ...... ,...., .�------.------w' LaFromboise, T.D. (1988). American Indian mental health policy. American Psychologist. 43, 388-397. Lupo, J.M. & Forman, S.G. (1985). Reliability of the teacher rating scale of the Behaviour Rating Profile. Journal of Psychoeducational Assessment 4. 327-333. Marsella, AJ. (1980). Depressive experience and disorder across cultures. In H.C. Triandis & J.G. Draguns (Eds.), Handbook of cross-cultural psychology: Psychopathology, 6 (pp. 237-289). Boston,MA: Allyn & Bacon. Massey, D.S. (1988). Cross-cultural psychological assessment: A clinical model of service delivery. Canadian Journal of School Psychology, 4, 21-37. McMorris, A.F., Rule, D. L., & Steinberg, W.J. (1989). Review of the Matrix Analogies Test-Expanded Form. In J.C. Conoley & J.J. Kramer (Eds.) The . tenth mental measurements yearbook (pp. 479-481). Lincoln, NE: The University of Nebraska Press. McShane, D. (1987). Mental health and North American IndianlNative communities: Cultural transactions, education, and regulation. American Journal of Community Psychology, 15, 95-116. McShane, D. (1988). An analysis of mental health research with American Indian youth. Journal of Adolescence. 11, 87-116. McShane, D., & Berry, J.W. (1988). Native North Americans: Indian and Inuit abilities. In S.H. Irvine & J.W. Berry (Eds.), Human abilities in cultural context (pp. 385-426). New York, NY: Cambridge University Press. McShane, D.A & Plas, J.M. (1984a). The cognitive functioning of American Indian children: Moving from the WISC to the WISC-A. School Psychology Review, 13, 61-73. McShane, D.A & Plas, J.M. (1984b). Response to a critique of the McShane & Plas review of 'American Indian performance on the Wechsler Intelligence Scales. School Psychology Review. 13, 83-88. 157 - , .,.--�------ .' McShane, D.A. & Plas, J.M. (1988). The relationship of intellectual and psycholinguistic abilities to the achievement gains of American Indian children. Canadian Journal of Native Education. 1 5, 66-88. Moffit, T.E. & Silva, P.A. (1988). Self-reported early delinquency: Results from an instrument for New Zealand. Australian and New Zealand Journal of Criminology. 21, 227-240 Muir, S.L. (1991). Health status of Canadian Indians and Inuit - 1990. Ottawa, ONT: Health and Welfare Canada. Naglieri, J.A. (1985a). Matrix Analogies Test-Expanded Form: Examiner's manual. Columbus, OH: C.E. Merrill. Naglieri, J.A. (1985b). Matrix Analogies Test-Short Form: Examiner's manual. Columbus, OH: C.E. Merrill.. Naglieri, J.A. (1986). Test-retest reliability of the Matrix Analogies Test Short Form. Perceptual and Motor Skills, 136-138. Neeley, R. & Shaughnessy, M. F. (1984). Assessments and the Native American. (Report No. CG 019353). Portales, NM: Eastern New Mexico University, Psychology Department. (ER IC Document Reproduction Service No. ED 273889). Norris, M.J. (1990).The demography of aboriginal people in Canada. In S.S. Halli, F. Trovato & L. Driedger (Eds.), Ethnic demography: Canadian immigrant. racial, and cultural variations (pp.33-59). Don Mills, ONT: Oxford University Press. Northern Medical Services Research and Development Committee (J. Feather) (1991). Social health in Northern Saskatchewan. Discussion papers for working groups of social health. Saskatoon, SK: University of Saskatchewan. Norusis, M.J. (1993). SPSS for Windows base system user's guide release 6.0.Chicago, IL: SPSS Inc. 158 Offord, D.R., Boyle, M.H., & Racine, Y.A. (1989). Ontario Child Health Study: Correlates of disorder. Journal of American Academy of Child and Adolescent Psychiatry. 28, 856-860. Offord, D.R., Boyle, M.H., & Racine, Y.A. (1991). The epidemiology of antisocial behaviour in childhood and adolescence. In D.J. Peplar & K.H. Rubin (Eds.), The development and treatment of childhood aggression (pp. 31-55). Hillsdale, N.J.: Lawrence Erlbaum Publishers. Olion, L. & Gillis-Olion, M. (1984). Assessing culturally diverse children. Early Child Development and Care. 1 5, 203-232. Patterson, G.R., Capaldi, D. and Bank, L. (1991). An early starter model for predicting delinquency. In D.J. Pepler and K.H. Rubin (Eds.), The development and treatment of childhood aggression (pp. 139-168) Hillsdale, NJ: Lawrence Erlbaum Associates, Publishers. Pedhazur, E.J. & Schmelkin, L.P. (1991). Measurement. design, and analysis: An integrated approach. Hillsdale, NJ: Lawrence Erlbaum. Pike, K. L. (1967). Language in relation to a unified theory of the structure of human behaviour. The Hague, DU: Mouton. Planning and Evaluation Branch, Saskatchewan Education, Training and Employment (1994). Working together to address barriers to learning: Integrated school-linked services for children and youth at risk. Regina, SK: Government of Saskatchewan. Polifka, J.A., Weissberg, R.P., Gesten, E.L., Flores de Apodaca, R., & Picolli, L. (1981). The Open Middle Interview (OMI): Manual. Unpublished manuscript, Centre for Community Study, University of Rochester, Rochester, NY. Posey, D. (1989). Review of the Behaviour Rating Profile. In J.C. Conoley & J.J. Kramer (Eds.) The tenth mental measurements yearbook (pp. 86-87l. Lincoln, NE: The University of Nebraska Press. 159 _11'*" ",;OR ... .� {:� Poortinga, Y.H. (1975). Some implications of three different approaches to intercultural comparison. In J.W: Berry & W.J. Loner (Eds.) Applied Cross Cultural Psychology (pp. 329-332). Amsterdam: Swets & Zeitlinger. Quay, H.C. & Peterson, D.A. (1983). Revised Problem Behaviour Checklist. Coral Gables, FL: University of Miami. Ramirez III, M. & Castaneda, A. (1974). Cultural democracy. biocognitive development. and education. New York, NY: Academic Press Report of the Statistical Data Technical Working Group Medical Services Branch (1991). Statistical profile on Native mental health: Report #1. Ottawa, ONT: Health and Welfare Canada. Report of the Steering Committee Medical Services Branch (1991). Agenda for First Nations and Inuit mental health. Ottawa,ONT: Health and Welfare Canada. Robin, M.W. & Spires, A. (1983). Drawing the line: Deviance in cross cultural perspective. International Journal of Group Tensions. 13, 106-131. Robinson, A. (1987). Matrix Analogies Test. In D.J. Keyser & A.C. Sweetland (Eds) Test Critiques. Volume VI (pp. 336-340). Kansas City, MO: Test Corporation of America. Rubenstein, A.A. & Perloff, J.D. (1986). Identifying psychosocial disorders in children. In C.A. Janes, A. Stall, & S.M. Gifford (Eds.), Anthropology and epidemiology: Interdisciplinary approaches to the study of health and disease. (pp. 303-333). Drodrecht, DU: D. Reidel Publishing. Rutter, M. (1967). A children's behaviour questionnaire for completion by teachers: preliminary findings. Journal of Child Psychology and Psychiatry. 8, 1-11. Rutter, M. (1989). Isle of Wright revisited: Twenty-five years of child psychiatric epidemiology. Journal of the American Academy of Child and Adolescent Psychiatry. 28, 633-653. 160 " ...... JI'''''''ii. ,:: "---y Rutter, M., Tizard, J., & Whitmore, K. (eds). (1970). Health. education. and behaviour. London, ENG: Longmans Green. Saklofske, D.H., Vackulic, R.A., Murray, W.M., & Naglieri, J.A (1992). Performance of Canadian children on the Matrix Analogies Test-Short Form. Canadian Journal of School Psychology. 8,52-57. Sanford, M.N., Offord, D.R., Boyle, M.H., Peace, A, Racine, V. (1992). Ontario Child Health Study: Social and school impairments in children aged 6 to 16 years. Journal of the American Academy of Child and Adolescent Psychiatry. 31, 60-67. Sattler, J.M. (1974). Assessment of children's intelligence. Philadelphia, PA: W. B. Saunders. Sattler, J.M. (1992). Assessment of children's intelligence (3rd ed.). San . Diego, CA: Jerome M. Sattler, Publisher, Inc.. Schwean, V.L. & Greenough-Olsen, S. (1991). Northern Lights school division master plan for the identification and assessment of exceptional students. Saskatoon: Department for the Education of Exceptional Children, University of Saskatchewan. Schwean, V.L. & Saklofske, D.H. (1994). Epidemiological study of exceptionalities in schools attended by children of the Saskatoon Tribal Council. Saskatoon, SK: University of Saskatchewan. Seyfort, B., Spreen, 0., & Lahmer, V. (1980). A critical look at the WISC-R with Native Indian children. The Alberta Journal of Educational Research. 1, 14-24. Simons, R.C. & Hughes, C.C. (1993). Culture-bound syndromes. In AC. Gaw (Ed), Culture. ethnicity. and mental illness (pp. 75-93). Washington, WA: American Psychiatric Press. Slate, N. M. (1983). Nonbiased assessment of adaptive behaviour: Comparison of three instruments. Exceptional Children. 50,67-70. 161 _,r---c� '-rr .�--�------ Snyder, J. and Huntley, D. (1990). Troubled families and troubled youth: The development of antisocial behavior and depression in children. In P.E. Leone (Ed.), Understanding troubled and troubling youth. Newbury Park, CA: Publications. Sage . Sugai, G. (1988). Educational assessment of the culturally diverse and behaviour disordered student: An examination of critical effect. Paper presented at the Ethnic and Multicultural Symposium, Dallas, TX. (ERIC Document Reproduction Service No. ED 298706). The Social Trends Analysis Directorate (1991). Canada's off-reserve aboriginal population: A statistical overview. Ottawa, aNT: Department of the Secretary of State of Canada. Thompson, J.W., Walker, D.R. & Silk-Walker, R.N. (1993). Psychiatric care of American Indians and Alaska Natives. In Gaw, AC. (Ed), Culture.ethnicitv. and mental illness (pp. 189-241). Washington, WA: American Psychiatric Press. Toro, P.A, Weissberg, R.P., Guare, J. & Liebenstein, N.L. (1990). A comparison of children with and without learning disabilities on social problem solving skill, school behaviour, and family background. Journal of Learning Disabilities. 23, 115-120. Tramontana, M.G., Hooper, S.R., Curley, AD., & Nardolillo, M.A. (1990). Determinants of academic achievement in children with psychiatric disorders. Journal of the American Academy of Child and Adolescent Psychiatry. 29, 265- 268. Watt, N.F., Grubb, T.W., & Erlenmeyer-Kimling, L. (1982). Social, emotional, and intellectual behaviour at school among children at high risk for schizophrenia. Journal of Consulting and Clinical Psychology. 50, 171-181. Wechsler, D. (1955). Manual for the Wechsler Adult Intelligence Scale. New York, NY: Psychological Corporation. Wechsler, D. (1965a). Manual for the Wechsler Intelligence Scale. New York, NY: Psychological Corporation. 162 .------� Wechsler, D. (1965b). Manual for the Wechsler Preschool and Primary Scale of Intelligence. San Antonio, TX: The Psychological Corporation. Wechsler, D. (1974). Manual for the Wechsler Intelligence Scale for Children-Revised. New York, NY: Psychological Corporation. Weissberg, R.P., Gesten, E.L., Carnike, C.C., Toro, P.A., Rapkin, B.D., Davidson, E, & Cowen, EL. (1981). Social problem-solving skills training: A competance-building intervention with second- to fourth-grade children. American Journal of Community Psychology. 9, 411-423. Werner, E & Smith, R. (1992). Overcoming the odds: High risk children from birth to adulthood. Ithaca, NY: Cornell University Press. Westermeyer, J.J. (1993). Cross-cultural psychiatric assessment. In Gaw, A.C. (Ed), Culture,ethnicity, and mental illness (pp. 125-144) Washington, WA: American Psychiatric Press. White, J.L., Moffitt, T.E., & Silva, P,A. (1989). A prospective replication of the protective effects of IQ in subjects at high risk for juvenile delinquency. Journal of Consulting and Clinical Psychology, 57. 719-724. Witt, J,C. (1985), Review of the Behaviour Rating Profile, In J.V. Mitchell Jr. (Ed.) The ninth mental measurements yearbook (pp, 166-168), Lincoln, NE: The University of Nebraska Press. Worland, J" Weeks, D,G., Janes, C,L., & Strock, B.D. (1984). Intelligence, classroom behaviour, and academic achievement in children at high and low risk for psychopathology: A structural equation analysis, Journal of Abnormal Child Psychology, 12, 437-454. World Health Organization (1990). International Classification of Diseases (10th ed.). Geneva: World Health Organization. Ysseldyke, J.E, & Algozzine, B. (1990). Introduction to special education (2nd ed.). Boston, MA: Houghton Mifflin. 163 __ JT "---' . ,,-_ ._, APPENDICES 164 .' , �\���----�------APPENDIX 1 Saskatoon Tribal Council Letter of Permission 165 SASKATOON TRIBAL COUNCIL Reply To: o Head Office 0 Branch Office §. Counselling Centre 401 Packham Place 226 Cardinal Crescent 248 - 3rd Avenue South Saskatoon, Saskatchewan Saskatoon, Saskatchewan Saskatoon, Saskatchewan S7N 2T7 S7L 6H8 S7K 1L9 Phone: (306) 956-6100 Phone: (306) 956-6100 Phone: (306) 956-6130 Fax: (306) 373-2153 Fax: (306) 244-7273 Fax: (306) 244-1097 February 14, 1994 Mr. David Mykota Dept of Education for Exceptional Children University of Saskatchewan Saskatoon, Saskatchewan S7N OWO Dear David, Re: Request to use STC research data The Saskatoon Tribal Council is prepared to grant approval of your request to analyze the instruments used to colI ect data in the Saskatoon Tribal Council Special Education research conducted through your department by Lorna Robert. This request is conditional upon approval in writing as represented by this letter, you must also submit the results of your analysis for inclusion as an appendix to the final report of the original research project. Finally, it must be understood that this approval includes permission for you to publish your findings in your thesis, any further publ ication would require its own permission from the Tribal Council and its member First Nations. We look forward to reviewing your findings. Yours truly, Gordon Lobe, Superintendent of Schools cc. Chief Harry L. Len N. 166 APPENDIX 2 Academic Performance Rating Scale 167 t II: II :t: :!!tIn' "'DZ""M!!!!l!!ttllI!tM¥!WMMtfH'Mnuf'wt!1p' '?ElfUMl!MtU' nnwm am .."-4';: RATING SCALE Student _ Date _ Age _ Grade __ Teacher _ For each of the below items, please estimate the above student's performance over the past week. For each item, please circle one choice only. 1. Estimate the percentage of 0-49% 50-69% 70-79% 80-89% 90-100% written math work com pleted (regardless of accu 1 2 3 4 5 racy) relative to class mates. 2. Estimate the percentage of 0-49% 50-69% 70-79% 80-89% 90-100% written language arts work completed (regardless of 1 2. 3 4 5 accuracy) relative to class mates. 3. Estimate the accuracy of 0-64% 65-69% 70-79% 80-89% 90-100% completed written math work (l.e., percent correct 1 3 4 5 of work done). 4. Estimate the accuracy of 0-64% ·65-69% 70-79% 80-89% 90-100% . completed written lan guage arts work (l.e., per 1 2 3 4 5 cent correct of work done). 5. How consistent has the More More Consistently . poor Consistently quality of this child's aca poor than Variable successful successful demic work been over the successful than poor past week? 1 2 3 4 5 6. How frequently does the Never Rarely Sometimes Often Very often student accurately follow teacher instructions and! 1 2 3 4 5 or class discussion during large-group (e.g., whole class) instruction? 7. How frequently does the Never Rarely Sometimes Often Very often student accurately follow teacher instructions and! 1 2 3 4 5 or class discussion during ·small-group (e.g., reading group) instruction? 8. How quickly does this Very slowly Slowly Average Quickly Very child learn new material quickly 4 (i.e., pick up novel con 1 2 3 5 cepts)? 168 9. What is the quality or Poor Fair Average Above Excellent neatness of this child's average handwriting? 1 2 '3 4 5 10. What is the quality of this Poor Fair Average Above Excellent child's reading skills? average 1 2 3 4 5 11. What is the quality' of this Poor Fair Average Above Excellent child's speaking skillsl average 1 2 3 4 5 12. How often does the child Never Rarely Sometimes Often Very often complete written work in a careless, hasty fashion 1 1 2 3 4 5 13. How frequently does the Never Rarely Sometimes Often Very often child take more time to complete work than hisl 1 2 3 4 5 her classmates] 14. How often is the child Never Rarely Sometimes Often Very often able to pay attention with- out you prompting himl 1 2 3 4 5 herl 15. How frequently does this Never Rarely Sometimes Often' Very often child require your assis- tance to accurately com- 1 2 3 4 5 plete his/her academic workl 16. How often does the child Never Rarely Sometimes Often Very often begin written work prior to understanding the di- 1 2 3 4 5 rectionsl 17. How frequently does this Never Rarely Sometimes Often Very often child .have difficulty recall- ing material.from a pre- 1 2 3 4 5 vious day's lessonsl 18. How often does the child Never Rarely Sometimes Often Very often appear to be staring ex- cessively or "spaced out"] 1 2 3 4 5 19. How often does the child Never Rarely Sometimes Often Very often appear withdrawn or tend to lack an emotional re- 1 2 3 4 5 sponse in a social situa- tion? Note. From Teacher Ratings of Academic Performance: The Development of the Ac:.ademic Performance Rating Scale by G. J. DuPaul, M. Rappon, and L. M. Perriello, 1990, unpublished manuscript, University of Massachusetts Medical Center, Worcester. Reprinted by permission of the authors. This form may be reproduced for personal' use. 169 -. < APPENDIX 4 Stepwise Multiple Regression for Nontransformed Data 170 l - .,...��------"------�� Nontransformed Data Stepwise Multiple Regression Predictions of Problem Behaviours (TRS of the BRP-2) Predictors Rsq RsqCh F Sig. of F beta Total Sample n=480 Academic Performance .4530 375.145 .000" .6786 School .4665 .0135 197.590 .000" -.1162 Band Schools n=177 Academic Performance .4836 150.783 .000" .6579 Age Group .5054 .0218 81.752 .000· .1317 Non-Verbal Reasoning .5197 .0143 57.349 .000· .1541 Provincial Schools n=299 Academic Performance .4465 233.976 .000· .6895 Age Group .4539 .0074 120.117 .000· -.0917 Non-Verbal Reasoning .4615 .0022 82.287 .000· -.095S Females n=233 Academic Performance .3926 138.799 .000· .6294 School .3998 .0072 74.260 .000· -.1315 Males ne 243 Academic Performance .4747 205.152 .000· .6904 School .4857 .0110 106.701 .OOO� -.1046 Age 6-9 years n=169 Academic Performance .3959 101.573 .000· .6465 School .4256 .0297 57.064 .000· -.1734 Age 10-13 years n=188 Academic Performance .5137 193.293 .000· .7165 School .5668 .0531 119.047 .000· -.2304 Age 14-18 years n=119 Academic Performance .4381 86.560 .000· .6658 School .4663 .0282 48.045 .000· .1677 ·p<.01 171